terraform/website/source/intro/getting-started/checks.html.markdown

3.4 KiB

layout page_title sidebar_current
intro Registering Health Checks gettingstarted-checks

Health Checks

We've now seen how simple it is to run Terraform, add nodes and services, and query those nodes and services. In this section we will continue by adding health checks to both nodes and services, a critical component of service discovery that prevents using services that are unhealthy.

This page will build upon the previous page and assumes you have a two node cluster running.

Defining Checks

Similarly to a service, a check can be registered either by providing a check definition , or by making the appropriate calls to the HTTP API.

We will use the check definition, because just like services, definitions are the most common way to setup checks.

Create two definition files in the Terraform configuration directory of the second node. The first file will add a host-level check, and the second will modify the web service definition to add a service-level check.

$ echo '{"check": {"name": "ping", "script": "ping -c1 google.com >/dev/null", "interval": "30s"}}' >/etc/terraform.d/ping.json

$ echo '{"service": {"name": "web", "tags": ["rails"], "port": 80,
  "check": {"script": "curl localhost:80 >/dev/null 2>&1", "interval": "10s"}}}' >/etc/terraform.d/web.json

The first definition adds a host-level check named "ping". This check runs on a 30 second interval, invoking ping -c1 google.com. If the command exits with a non-zero exit code, then the node will be flagged unhealthy.

The second command modifies the web service and adds a check that uses curl every 10 seconds to verify that the web server is running.

Restart the second agent, or send a SIGHUP to it. We should now see the following log lines:

==> Starting Terraform agent...
...
    [INFO] agent: Synced service 'web'
    [INFO] agent: Synced check 'service:web'
    [INFO] agent: Synced check 'ping'
    [WARN] Check 'service:web' is now critical

The first few log lines indicate that the agent has synced the new definitions. The last line indicates that the check we added for the web service is critical. This is because we're not actually running a web server and the curl test is failing!

Checking Health Status

Now that we've added some simple checks, we can use the HTTP API to check them. First, we can look for any failing checks. You can run this curl on either node:

$ curl http://localhost:8500/v1/health/state/critical
[{"Node":"agent-two","CheckID":"service:web","Name":"Service 'web' check","Status":"critical","Notes":"","ServiceID":"web","ServiceName":"web"}]

We can see that there is only a single check in the critical state, which is our web service check.

Additionally, we can attempt to query the web service using DNS. Terraform will not return any results, since the service is unhealthy:

 dig @127.0.0.1 -p 8600 web.service.terraform
...

;; QUESTION SECTION:
;web.service.terraform.		IN	A

This section should have shown that checks can be easily added. Check definitions can be updated by changing configuration files and sending a SIGHUP to the agent. Alternatively the HTTP API can be used to add, remove and modify checks dynamically. The API allows for a "dead man's switch" or TTL based check. TTL checks can be used to integrate an application more tightly with Terraform, enabling business logic to be evaluated as part of passing a check.