istio.io/content/docs/concepts/traffic-management/index.md

1042 lines
37 KiB
Markdown

---
title: Traffic Management
description: Describes the various Istio features focused on traffic routing and control.
weight: 20
keywords: [traffic-management]
aliases:
- /docs/concepts/traffic-management/overview
- /docs/concepts/traffic-management/pilot
- /docs/concepts/traffic-management/rules-configuration
- /docs/concepts/traffic-management/fault-injection
- /docs/concepts/traffic-management/handling-failures
- /docs/concepts/traffic-management/load-balancing
- /docs/concepts/traffic-management/request-routing
---
This page provides an overview of how traffic management works
in Istio, including the benefits of its traffic management
principles. It assumes that you've already read [What is Istio?](/docs/concepts/what-is-istio/)
and are familiar with Istio's high-level architecture.
Using Istio's traffic management model essentially decouples traffic flow
and infrastructure scaling, letting you specify via Pilot what
rules they want traffic to follow rather than which specific pods/VMs should
receive traffic - Pilot and intelligent Envoy proxies look after the
rest. For example, you can specify via Pilot that you want 5%
of traffic for a particular service to go to a canary version irrespective
of the size of the canary deployment, or send traffic to a particular version
depending on the content of the request.
{{< image width="85%" ratio="69.52%"
link="./TrafficManagementOverview.svg"
caption="Traffic Management with Istio"
>}}
Decoupling traffic flow from infrastructure scaling allows Istio
to provide a variety of traffic management features that live outside the
application code. As well as dynamic [request routing](#request-routing)
for A/B testing, gradual rollouts, and canary releases, it also handles
[failure recovery](#handling-failures) using timeouts, retries,
circuit breakers, and [fault injection](#fault-injection) to
test the compatibility of failure recovery policies across services. These
capabilities are all realized through the Envoy sidecars/proxies deployed
across the service mesh.
## Pilot and Envoy
The core component used for traffic management in Istio is **Pilot**, which
manages and configures all the Envoy
proxy instances deployed in a particular Istio service mesh. Pilot lets you
specify which rules you want to use to route traffic between Envoy proxies
and configure failure recovery features such as timeouts, retries, and
circuit breakers. It also maintains a canonical model of all the services
in the mesh and uses this model to let Envoy instances know about the other Envoy instances in the mesh via its discovery service.
Each Envoy instance maintains [load balancing information](#discovery-and-load-balancing)
based on the information it gets from Pilot and periodic health-checks
of other instances in its load-balancing pool, allowing it to intelligently
distribute traffic between destination instances while following its specified
routing rules.
Pilot is responsible for the lifecycle of Envoy instances deployed
across the Istio service mesh.
{{< image width="60%" ratio="72.17%"
link="./PilotAdapters.svg"
caption="Pilot Architecture"
>}}
As shown in the figure above, Pilot maintains a canonical
representation of services in the mesh that is independent of the underlying
platform. Platform-specific adapters in Pilot are responsible for
populating this canonical model appropriately. For example, the Kubernetes
adapter in Pilot implements the necessary controllers to watch the
Kubernetes API server for changes to the pod registration information, ingress
resources, and third-party resources that store traffic management rules.
This data is translated into the canonical representation. An Envoy-specific configuration is then generated based on the canonical representation.
Pilot enables [service discovery](https://www.envoyproxy.io/docs/envoy/latest/api-v1/cluster_manager/sds),
dynamic updates to [load balancing pools](https://www.envoyproxy.io/docs/envoy/latest/configuration/cluster_manager/cds)
and [routing tables](https://www.envoyproxy.io/docs/envoy/latest/configuration/http_conn_man/rds).
You can specify high-level traffic management rules through
[Pilot's Rule configuration](/docs/reference/config/istio.networking.v1alpha3/). These rules are translated into low-level
configurations and distributed to Envoy instances.
## Request routing
As described above, the canonical representation
of services in a mesh is maintained by Pilot. The Istio
model of a service is independent of how it is represented in the underlying
platform (Kubernetes, Mesos, Cloud Foundry,
etc.). Platform-specific adapters are responsible for populating the
internal model representation with various fields from the metadata found
in the platform.
Istio introduces the concept of a _service version_, which is a finer-grained
way to subdivide service instances by versions (`v1`, `v2`) or environment
(`staging`, `prod`). These variants are not necessarily different API
versions: they could be iterative changes to the same service, deployed in
different environments (prod, staging, dev, etc.). Common scenarios where
this is used include A/B testing or canary rollouts. Istio's [traffic
routing rules](#rule-configuration) can refer to service versions to provide
additional control over traffic between services.
### Communication between services
{{< image width="60%" ratio="100.42%"
link="./ServiceModel_Versions.svg"
alt="Showing how service versions are handled."
caption="Service Versions"
>}}
As shown in the figure above, clients of a service have no knowledge
of different versions of the service. They can continue to access the
services using the hostname/IP address of the service. The Envoy sidecar/proxy
intercepts and forwards all requests/responses between the client and the
service.
Envoy determines its actual choice of service version dynamically
based on the routing rules that you specify by using Pilot. This
model enables the application code to decouple itself from the evolution of its dependent
services, while providing other benefits as well (see
[Mixer](/docs/concepts/policies-and-telemetry/)). Routing
rules allow Envoy to select a version based
on conditions such as headers, tags associated with
source/destination, and/or by weights assigned to each version.
Istio also provides load balancing for traffic to multiple instances of
the same service version. See [Discovery
and Load Balancing](/docs/concepts/traffic-management/#discovery-and-load-balancing) for more.
Istio does not provide a DNS. Applications can try to resolve the
FQDN using the DNS service present in the underlying platform (`kube-dns`,
`mesos-dns`, etc.).
### Ingress and egress
Istio assumes that all traffic entering and leaving the service mesh
transits through Envoy proxies. By deploying an Envoy proxy in front of
services, you can conduct A/B testing, deploy canary services,
etc. for user-facing services. Similarly, by routing traffic to external
web services (for instance, accessing a maps API or a video service API)
via the Envoy sidecar, you can add failure recovery features such as
timeouts, retries, and circuit breakers and obtain detailed metrics on
the connections to these services.
{{< image width="60%" ratio="28.88%"
link="./ServiceModel_RequestFlow.svg"
alt="Ingress and Egress through Envoy."
caption="Request Flow"
>}}
## Discovery and load balancing
Istio load balances traffic across instances of a service in a service mesh.
Istio assumes the presence of a service registry
to keep track of the pods/VMs of a service in the application. It also
assumes that new instances of a service are automatically registered with
the service registry and unhealthy instances are automatically removed.
Platforms such as Kubernetes and Mesos already provide such functionality for
container-based applications, and many solutions exist for VM-based
applications.
Pilot consumes information from the service
registry and provides a platform-independent service discovery
interface. Envoy instances in the mesh perform service discovery and
dynamically update their load balancing pools accordingly.
{{< image width="55%" ratio="74.79%"
link="./LoadBalancing.svg"
caption="Discovery and Load Balancing"
>}}
As shown in the figure above, services in the mesh access each other
using their DNS names. All HTTP traffic bound to a service is automatically
re-routed through Envoy. Envoy distributes the traffic across instances in
the load balancing pool. While Envoy supports several
[sophisticated load balancing algorithms](https://www.envoyproxy.io/docs/envoy/latest/intro/arch_overview/load_balancing),
Istio currently allows three load balancing modes:
round robin, random, and weighted least request.
In addition to load balancing, Envoy periodically checks the health of each
instance in the pool. Envoy follows a circuit breaker pattern to
classify instances as unhealthy or healthy based on their failure rates for
the health check API call. In other words, when the number of health
check failures for a given instance exceeds a pre-specified threshold, it
will be ejected from the load balancing pool. Similarly, when the number of
health checks that pass exceed a pre-specified threshold, the instance will
be added back into the load balancing pool. You can find out more about Envoy's
failure-handling features in [Handling Failures](#handling-failures).
Services can actively shed load by responding with an HTTP 503 to a health
check. In such an event, the service instance will be immediately removed
from the caller's load balancing pool.
## Handling failures
Envoy provides a set of out-of-the-box _opt-in_ failure recovery features
that can be taken advantage of by the services in an application. Features
include:
1. Timeouts
1. Bounded retries with timeout budgets and variable jitter between retries
1. Limits on number of concurrent connections and requests to upstream services
1. Active (periodic) health checks on each member of the load balancing pool
1. Fine-grained circuit breakers (passive health checks) -- applied per
instance in the load balancing pool
These features can be dynamically configured at runtime through
[Istio's traffic management rules](#rule-configuration).
The jitter between retries minimizes the impact of retries on an overloaded
upstream service, while timeout budgets ensure that the calling service
gets a response (success/failure) within a predictable time frame.
A combination of active and passive health checks (4 and 5 above)
minimize the chances of accessing an unhealthy instance in the load
balancing pool. When combined with platform-level health checks (such as
those supported by Kubernetes or Mesos), applications can ensure that
unhealthy pods/containers/VMs can be quickly ejected from the service
mesh, minimizing the request failures and impact on latency.
Together, these features enable the service mesh to tolerate failing nodes
and prevent localized failures from cascading instability to other nodes.
### Fine tuning
Istio's traffic management rules allow you to set global defaults for failure recovery per
service/version. However, consumers of a service can also override
[timeout](/docs/reference/config/istio.networking.v1alpha3/#HTTPRoute-timeout)
and
[retry](/docs/reference/config/istio.networking.v1alpha3/#HTTPRoute-retries)
defaults by providing request-level overrides through special HTTP headers.
With the Envoy proxy implementation, the headers are `x-envoy-upstream-rq-timeout-ms` and
`x-envoy-max-retries`, respectively.
### Failure handling FAQ
Q: *Do applications still handle failures when running in Istio?*
Yes. Istio improves the reliability and availability of services in the
mesh. However, **applications need to handle the failure (errors)
and take appropriate fallback actions**. For example, when all instances in
a load balancing pool have failed, Envoy will return HTTP 503. It is the
responsibility of the application to implement any fallback logic that is
needed to handle the HTTP 503 error code from an upstream service.
Q: *Will Envoy's failure recovery features break applications that already
use fault tolerance libraries (for example [Hystrix](https://github.com/Netflix/Hystrix))?*
No. Envoy is completely transparent to the application. A failure response
returned by Envoy would not be distinguishable from a failure response
returned by the upstream service to which the call was made.
Q: *How will failures be handled when using application-level libraries and
Envoy at the same time?*
Given two failure recovery policies for the same destination service, **the
more restrictive of the two will be triggered when failures occur**. For example, you have two timeouts -- one set in Envoy and another in an application's library. In this
example, if the application sets a 5 second timeout for an API call to a
service, while the you configured a 10 second timeout in Envoy, the
application's timeout will kick in first. Similarly, if Envoy's circuit
breaker triggers before the application's circuit breaker, API calls to the
service will get a 503 from Envoy.
## Fault injection
While the Envoy sidecar/proxy provides a host of
[failure recovery mechanisms](#handling-failures) to services running
on Istio, it is still
imperative to test the end-to-end failure recovery capability of the
application as a whole. Misconfigured failure recovery policies (for example,
incompatible/restrictive timeouts across service calls) could result in
continued unavailability of critical services in the application, resulting
in poor user experience.
Istio enables protocol-specific fault injection into the network, instead
of killing pods or delaying or corrupting packets at the TCP layer. The rationale
is that the failures observed by the application layer are the same
regardless of network level failures, and that more meaningful failures can
be injected at the application layer (for example, HTTP error codes) to exercise the resilience of an application.
You can configure faults to be injected into requests that match
specific conditions. You can further restrict the percentage of
requests that should be subjected to faults. Two types of faults can be
injected: delays and aborts. Delays are timing failures, mimicking
increased network latency, or an overloaded upstream service. Aborts are
crash failures that mimic failures in upstream services. Aborts usually
manifest in the form of HTTP error codes or TCP connection failures.
## Rule configuration
Istio provides a simple configuration model to
control how API calls and layer-4 traffic flow across various
services in an application deployment. The configuration model allows you to
configure service-level properties such as circuit breakers, timeouts,
and retries, as well as set up common continuous deployment tasks such as
canary rollouts, A/B testing, staged rollouts with %-based traffic splits,
etc.
There are four traffic management configuration resources in Istio:
**VirtualService**, **DestinationRule**, **ServiceEntry**, and **Gateway**:
* A [VirtualService](/docs/reference/config/istio.networking.v1alpha3/#VirtualService)
defines the rules that control how requests for a service are routed within an Istio service mesh.
* A [DestinationRule](/docs/reference/config/istio.networking.v1alpha3/#DestinationRule)
configures the set of policies to be applied to a request after `VirtualService` routing has occurred.
* A [ServiceEntry](/docs/reference/config/istio.networking.v1alpha3/#ServiceEntry) is commonly used to enable requests to services outside of an Istio service mesh.
* A [Gateway](/docs/reference/config/istio.networking.v1alpha3/#Gateway)
configures a load balancer for HTTP/TCP traffic, most commonly operating at the edge of the mesh to enable ingress traffic for an application.
The following sections describe these resources. See [networking reference](/docs/reference/config/istio.networking.v1alpha3/)
for detailed reference information.
For example, you can implement a simple rule to send 100% of incoming traffic for a *reviews* service to version "v1" by using a `VirtualService` configuration as follows:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: reviews
spec:
hosts:
- reviews
http:
- route:
- destination:
host: reviews
subset: v1
{{< /text >}}
This configuration says that traffic sent to the *reviews* service
(specified in the `hosts` field) should be routed to the v1 subset
of the underlying *reviews* service instances. The route `subset` specifies the name of a defined subset in a corresponding destination rule configuration.
A subset specifies one or more labels that identify version-specific instances.
For example, in a Kubernetes deployment of Istio, "version: v1" indicates that
only pods containing the label "version: v1" will receive traffic.
In a `DestinationRule`, you can then add additional policies. For example, the following definition specifies to use the random load balancing mode:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: DestinationRule
metadata:
name: reviews
spec:
host: reviews
trafficPolicy:
loadBalancer:
simple: RANDOM
subsets:
- name: v1
labels:
version: v1
- name: v2
labels:
version: v2
{{< /text >}}
Rules can be configured using the
[istioctl CLI](/docs/reference/commands/istioctl/), or in a Kubernetes
deployment using the `kubectl` command instead, although only `istioctl` will
perform model validation and is recommended. See the
[configuring request routing task](/docs/tasks/traffic-management/request-routing/)
for examples.
### Virtual Services
A [VirtualService](/docs/reference/config/istio.networking.v1alpha3/#VirtualService)
defines the rules that control how requests for a service are routed within an Istio service mesh.
For example, a virtual service could route requests to different versions of a service or to a completely different service than was requested.
Requests can be routed based on the request source and destination, HTTP paths and
header fields, and weights associated with individual service versions.
#### Rule destinations
Routing rules correspond to one or more request destination hosts that are specified in
a `VirtualService` configuration. These hosts may or may not be the same as the actual
destination workload and may not even correspond to an actual routable service in the mesh.
For example, to define routing rules for requests to the *reviews* service using its internal
mesh name `reviews` or via host `bookinfo.com`, a `VirtualService` could set the `hosts` field as:
{{< text yaml >}}
hosts:
- reviews
- bookinfo.com
{{< /text >}}
The `hosts` field specifies, implicitly or explicitly, one or more fully qualified
domain names (FQDN). The short name `reviews`, above, would implicitly
expand to an implementation specific FQDN. For example, in a Kubernetes environment
the full name is derived from the cluster and namespace of the `VirtualSevice`
(for example, `reviews.default.svc.cluster.local`).
#### Splitting traffic between versions
Each route rule identifies one or more weighted backends to call when the rule is activated.
Each backend corresponds to a specific version of the destination service,
where versions can be expressed using _labels_.
If there are multiple registered instances with the specified label(s),
they will be routed to based on the load balancing policy configured for the service,
or round-robin by default.
For example, the following rule will route 25% of traffic for the *reviews* service to instances with
the "v2" label and the remaining 75% of traffic to "v1":
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: reviews
spec:
hosts:
- reviews
http:
- route:
- destination:
host: reviews
subset: v1
weight: 75
- destination:
host: reviews
subset: v2
weight: 25
{{< /text >}}
#### Timeouts and retries
By default, the timeout for HTTP requests is 15 seconds,
but it can be overridden in a route rule as follows:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: ratings
spec:
hosts:
- ratings
http:
- route:
- destination:
host: ratings
subset: v1
timeout: 10s
{{< /text >}}
You can also specify the number of retry attempts for an HTTP request in a route rule.
The maximum number of retry attempts, or the number of attempts possible within the default or overridden timeout period, can be set as follows:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: ratings
spec:
hosts:
- ratings
http:
- route:
- destination:
host: ratings
subset: v1
retries:
attempts: 3
perTryTimeout: 2s
{{< /text >}}
Note that request timeouts and retries can also be
[overridden on a per-request basis](#fine-tuning).
See the [request timeouts task](/docs/tasks/traffic-management/request-timeouts) for an example of timeout control.
#### Injecting faults
A route rule can specify one or more faults to inject
while forwarding HTTP requests to the rule's corresponding request destination.
The faults can be either delays or aborts.
The following example introduces a 5 second delay in 10% of the requests to the "v1" version of the *ratings* microservice:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: ratings
spec:
hosts:
- ratings
http:
- fault:
delay:
percent: 10
fixedDelay: 5s
route:
- destination:
host: ratings
subset: v1
{{< /text >}}
You can use the other kind of fault, an abort, to prematurely terminate a request. For example, to simulate a failure.
The following example returns an HTTP 400 error code for 10%
of the requests to the *ratings* service "v1":
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: ratings
spec:
hosts:
- ratings
http:
- fault:
abort:
percent: 10
httpStatus: 400
route:
- destination:
host: ratings
subset: v1
{{< /text >}}
Sometimes delay and abort faults are used together. For example, the following rule delays by 5 seconds all requests from the *reviews* service "v2" to the *ratings* service "v1" and then aborts 10% of them:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: ratings
spec:
hosts:
- ratings
http:
- match:
- sourceLabels:
app: reviews
version: v2
fault:
delay:
fixedDelay: 5s
abort:
percent: 10
httpStatus: 400
route:
- destination:
host: ratings
subset: v1
{{< /text >}}
To see fault injection in action, see the [fault injection task](/docs/tasks/traffic-management/fault-injection/).
#### Conditional rules
Rules can optionally be qualified to only apply to requests that match some
specific condition such as the following:
_1. Restrict to specific client workloads using workload labels_. For example, a rule
can indicate that it only applies to calls from workloads (pods) implementing
the *reviews* service:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: ratings
spec:
hosts:
- ratings
http:
- match:
sourceLabels:
app: reviews
...
{{< /text >}}
The value of `sourceLabels` depends on the implementation of the service.
In Kubernetes, for example, it would probably be the same labels that are used
in the pod selector of the corresponding Kubernetes service.
The above example can also be further refined to only apply to calls from version "v2"
of the *reviews* service:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: ratings
spec:
hosts:
- ratings
http:
- match:
- sourceLabels:
app: reviews
version: v2
...
{{< /text >}}
_2. Select rule based on HTTP headers_. For example, the following rule only applies to an incoming request if it includes a "cookie" header that
contains the substring "user=jason":
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: reviews
spec:
hosts:
- reviews
http:
- match:
- headers:
cookie:
regex: "^(.*?;)?(user=jason)(;.*)?$"
...
{{< /text >}}
If more than one header is specified in the rule, then all of the
corresponding headers must match for the rule to apply.
_3. Select rule based on request URI_. For example, the following rule only applies to a request if the URI path starts with `/api/v1`:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: productpage
spec:
hosts:
- productpage
http:
- match:
- uri:
prefix: /api/v1
...
{{< /text >}}
#### Multiple match conditions
Multiple match conditions can be set simultaneously. In such a case, AND or OR
semantics apply, depending on the nesting.
If multiple conditions are nested in a single match clause, then the conditions
are ANDed. For example, the following rule only applies if the
client workload is “reviews:v2” AND the "cookie" header containing
"user=jason" is present in the request:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: ratings
spec:
hosts:
- ratings
http:
- match:
- sourceLabels:
app: reviews
version: v2
headers:
cookie:
regex: "^(.*?;)?(user=jason)(;.*)?$"
...
{{< /text >}}
If instead, the condition appear in separate match clauses, then only one
of the conditions applies (OR semantics):
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: ratings
spec:
hosts:
- ratings
http:
- match:
- sourceLabels:
app: reviews
version: v2
- headers:
cookie:
regex: "^(.*?;)?(user=jason)(;.*)?$"
...
{{< /text >}}
This rule applies if either the client workload is “reviews:v2” OR
the "cookie" header containing "user=jason" is present in the request.
#### Precedence
When there are multiple rules for a given destination,
they are evaluated in the order they appear
in the `VirtualService`, meaning the first rule
in the list has the highest priority.
**Why is priority important?** Whenever the routing story for a particular
service is purely weight based, it can be specified in a single rule.
On the other hand, when other conditions
(such as requests from a specific user) are being used to route traffic, more
than one rule will be needed to specify the routing. This is where the
rule priority must be carefully considered to make sure that the rules are
evaluated in the right order.
A common pattern for generalized route specification is to provide one or
more higher priority rules that match various conditions,
and then provide a single weight-based rule with no match
condition last to provide the weighted distribution of
traffic for all other cases.
For example, the following `VirtualService` contains two rules that, together,
specify that all requests for the *reviews* service that includes a header
named "Foo" with the value "bar" will be sent to the "v2" instances.
All remaining requests will be sent to "v1":
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: reviews
spec:
hosts:
- reviews
http:
- match:
- headers:
Foo:
exact: bar
route:
- destination:
host: reviews
subset: v2
- route:
- destination:
host: reviews
subset: v1
{{< /text >}}
Notice that the header-based rule has the higher priority. If
it was lower, these rules wouldn't work as expected because the weight-based
rule, with no specific match condition, would be evaluated first to route all traffic to "v1", even requests that include the
matching "Foo" header. Once a rule is found that applies to the incoming
request, it is executed and the rule-evaluation process terminates. That's why it's very important to carefully consider the
priorities of each rule when there is more than one.
### Destination rules
A [DestinationRule](/docs/reference/config/istio.networking.v1alpha3/#DestinationRule)
configures the set of policies to be applied to a request after `VirtualService` routing has occurred. They are
intended to be authored by service owners, describing the circuit breakers, load balancer settings, TLS settings, an other settings.
A `DestinationRule` also defines addressable `subsets`, meaning named versions, of the corresponding destination host.
These subsets are used in `VirtualService` route specifications when sending traffic to specific versions of the service.
The following `DestinationRule` configures policies and subsets for the reviews service:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: DestinationRule
metadata:
name: reviews
spec:
host: reviews
trafficPolicy:
loadBalancer:
simple: RANDOM
subsets:
- name: v1
labels:
version: v1
- name: v2
labels:
version: v2
trafficPolicy:
loadBalancer:
simple: ROUND_ROBIN
- name: v3
labels:
version: v3
{{< /text >}}
Notice that multiple policies, default and v2-specific in this example, can be
specified in a single `DestinationRule` configuration.
#### Circuit breakers
A simple circuit breaker can be set based on a number of conditions such as connection and request limits.
For example, the following `DestinationRule`
sets a limit of 100 connections to *reviews* service version "v1" backends:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: DestinationRule
metadata:
name: reviews
spec:
host: reviews
subsets:
- name: v1
labels:
version: v1
trafficPolicy:
connectionPool:
tcp:
maxConnections: 100
{{< /text >}}
See the [circuit-breaking task](/docs/tasks/traffic-management/circuit-breaking/) for a demonstration of circuit breaker control.
#### Rule evaluation
Similar to route rules, policies defined in a `DestinationRule` are associated with a particular *host*. However if they are subset specific,
activation depends on route rule evaluation results.
The first step in the rule evaluation process evaluates the route rules in
the `VirtualService` corresponding to the requested *host*, if there are any,
to determine the subset (meaning specific
version) of the destination service that the current request will be routed
to. Next, the set of policies corresponding to the selected subset, if any,
are evaluated to determine if they apply.
**NOTE:** One subtlety of the algorithm to keep in mind is that policies
that are defined for specific subsets will only be applied if
the corresponding subset is explicitly routed to. For example,
consider the following configuration as the one and only rule defined for the
*reviews* service, meaning there are no route rules in the corresponding `VirtualService` definition:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: DestinationRule
metadata:
name: reviews
spec:
host: reviews
subsets:
- name: v1
labels:
version: v1
trafficPolicy:
connectionPool:
tcp:
maxConnections: 100
{{< /text >}}
Since there is no specific route rule defined for the *reviews*
service, default round-robin routing behavior will apply, which will
presumably call "v1" instances on occasion, maybe even always if "v1" is
the only running version. Nevertheless, the above policy will never be
invoked since the default routing is done at a lower level. The rule
evaluation engine will be unaware of the final destination and therefore
unable to match the subset policy to the request.
You can fix the above example in one of two ways. You can either move the
traffic policy up a level in the `DestinationRule` to make it apply to any version:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: DestinationRule
metadata:
name: reviews
spec:
host: reviews
trafficPolicy:
connectionPool:
tcp:
maxConnections: 100
subsets:
- name: v1
labels:
version: v1
{{< /text >}}
Or, better yet, define proper route rules for the service in the `VirtualService` definition.
For example, add a simple route rule for "reviews:v1":
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: reviews
spec:
hosts:
- reviews
http:
- route:
- destination:
host: reviews
subset: v1
{{< /text >}}
Although the default Istio behavior conveniently sends traffic from any
source to all versions of a destination service
without any rules being set, as soon as version discrimination is desired
rules are going to be needed.
Therefore, setting a default rule for every service, right from the
start, is generally considered a best practice in Istio.
### Service entries
A [ServiceEntry](/docs/reference/config/istio.networking.v1alpha3/#ServiceEntry)
is used to add additional entries into the service registry that Istio maintains internally.
It is most commonly used to enable requests to services outside of an Istio service mesh.
For example, the following `ServiceEntry` can be used to allow external calls to services hosted
under the `*.foo.com` domain:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: ServiceEntry
metadata:
name: foo-ext-svc
spec:
hosts:
- *.foo.com
ports:
- number: 80
name: http
protocol: HTTP
- number: 443
name: https
protocol: HTTPS
{{< /text >}}
The destination of a `ServiceEntry` is specified using the `hosts` field, which
can be either a fully qualified or wildcard domain name.
It represents a white listed set of one or more services that services
in the mesh are allowed to access.
A `ServiceEntry` is not limited to external service configuration. It can be of two types: mesh-internal or mesh-external.
Mesh-internal entries are like all other internal services but are used to explicitly add services
to the mesh. They can be used to add services as part of expanding the service mesh to include unmanaged infrastructure
(for example, VMs added to a Kubernetes-based service mesh).
Mesh-external entries represent services external to the mesh.
For them, mutual TLS authentication is disabled and policy enforcement is performed on the client-side,
instead of on the server-side as it is for internal service requests.
Service entries work well in conjunction with virtual services
and destination rules as long as they refer to the services using matching
`hosts`. For example, the following rule can be used in conjunction with
the above `ServiceEntry` rule to set a 10s timeout for calls to
the external service at `bar.foo.com`:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: bar-foo-ext-svc
spec:
hosts:
- bar.foo.com
http:
- route:
- destination:
host: bar.foo.com
timeout: 10s
{{< /text >}}
Rules to redirect and forward traffic, to define retry,
timeout, and fault injection policies are all supported for external destinations.
Weighted (version-based) routing is not possible, however, since there is no notion
of multiple versions of an external service.
See the [egress task](/docs/tasks/traffic-management/egress/) for a more
about accessing external services.
### Gateways
A [Gateway](/docs/reference/config/istio.networking.v1alpha3/#Gateway)
configures a load balancer for HTTP/TCP traffic, most commonly operating at the edge of the
mesh to enable ingress traffic for an application.
Unlike Kubernetes Ingress, Istio `Gateway` only configures the L4-L6 functions
(for example, ports to expose, TLS configuration). Users can then use standard Istio rules
to control HTTP requests as well as TCP traffic entering a `Gateway` by binding a
`VirtualService` to it.
For example, the following simple `Gateway` configures a load balancer
to allow external HTTPS traffic for host `bookinfo.com` into the mesh:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: Gateway
metadata:
name: bookinfo-gateway
spec:
servers:
- port:
number: 443
name: https
protocol: HTTPS
hosts:
- bookinfo.com
tls:
mode: SIMPLE
serverCertificate: /tmp/tls.crt
privateKey: /tmp/tls.key
{{< /text >}}
To configure the corresponding routes, you must define a `VirtualService`
for the same host and bound to the `Gateway` using
the `gateways` field in the configuration:
{{< text yaml >}}
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
name: bookinfo
spec:
hosts:
- bookinfo.com
gateways:
- bookinfo-gateway # <---- bind to gateway
http:
- match:
- uri:
prefix: /reviews
route:
...
{{< /text >}}
See the [ingress task](/docs/tasks/traffic-management/ingress/) for a
complete ingress gateway example.
Although primarily used to manage ingress traffic, a `Gateway` can also be used to model
a purely internal or egress proxy. Irrespective of the location, all gateways
can be configured and controlled in the same way. See
[gateway reference](/docs/reference/config/istio.networking.v1alpha3/#Gateway)
for details.