Ability to define more than one health check per service

[petemounce]

I'd love to be able to do this. We run in AWS, with ELBs, and so each of our components will have at least one endpoint for the ELB-based health-check. This is a "I'm present" check.

Some of our more sophisticated components are starting to have more health checks, so we can assert that things are healthy in production. For example, pinging http dependencies, select'ing from a database, etc. It would be great if consul can ping these individually, so we can see this level of granularity via it.

Some of these checks indicate "cannot function"; some of them indicate "can function but degraded" - I think consul's 0=good, 1=warning, 2=critical has us covered here (thanks!).

(I've only tried against 0.4.1 - #367 alludes to something like this being possible already...?)

[wolfspyre]

+1 This would be helpful. I wonder about the complications that this would present around 'if check A fails, do I continue with check B?'

[petemounce]

Personally, while a dependsOn element in the checks would be quite expressive and I could make use of it, I could live without it.

[armon]

I think we have an open ticket for this already! This is a goal for 0.5. The core supports it already (internal catalog / APIs), but we need to expose it out.

[petemounce]

@armon that's great; looking forward to it landing!

[ryanuber]

@petemounce there is a PR open for this on #591, and an existing issue on #230. Closing this out to track in those issues, thanks!

[petemounce]

Brilliant!

Sent from my phone. Please excuse typos and brevity, but never text speak.
On 10 Jan 2015 00:16, "Ryan Uber" notifications@github.com wrote:

@petemounce https://github.com/petemounce there is a PR open for this
on #591 #591, and an existing
issue on #230 #230. Closing
this out to track in those issues, thanks!

—
Reply to this email directly or view it on GitHub
#562 (comment).