Are there any best practices documented for doing ...
# support-questions-legacy
n
Are there any best practices documented for doing Liveness probes when running self-hosted SuperTokens within your kubernetes cluster? Right now we do a HTTP call to
/hello
but the probe failes a few times a day with the folloing error:
Copy code
Liveness probe failed: Get "http://10.244.1.77:3567/hello": context deadline exceeded (Client.Timeout exceeded while awaiting headers)
r
hey!
this is the correct way. How often are you querying it? What is the server pool size config? And how much is the response timeout set to when querying it?
n
Note that the upper might be related to out cluster and not something you can solve btw 😄
r
We do a similar check for out managed service instances and we run these /hello queries every min on each instance. No problem like this has happened for us
So most likely some other issue with your setup
n
yeah, we avoided upgrading kubernetes for a while - maybe it is time 😄
okay, good to hear that hello is the "recommended" way of doing these checks, it would be nice if this was documented
r
Fair enough. Will add that to our todos. Or you can even contribute it to our docs 😇
Oh and by the way, a hello api call also makes sure that the core is connected to the db correctly. So if the db is down, hello will fail as well
Or if there is an issue between the core and db’s connection
n
That is great, Ill send a PR 🙂
r
Should probably add it in the page that talks about running with docker
r
Yup.
n
and also the db connection hanging/timping out will most likely be the issue on our side! Thank you again for these infos!
r
thanks for the PR! Will check it out soon