You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Enhance /api/status to return multiple status codes or more information
Problem
Today the status endpoint returns 200 or 500 and if you use this endpoint as a health check for a LB or other node manager, the node may get killed prematurely.
A 500 generally indicates one of our metrics has exceeded a prescribed threshold, most commonly a db connection wait, or similar issue.
This can simply mean the box is under heavy load, and not necessarily that the box is sick.
Solution
Either provide some configuration as to what the status endpoint returns or reserve 500 for a more catastrophic error - and provide some varying degrees of status so it functions better as a health check for a LB or other node manager.
Alternatives/workarounds
If the /api/status endpoint is causing you trouble as a health check endpoint, you can also make an API call such as /api/login with a known user + password w/out an applicationId - as a health check.
Please give us a thumbs up or thumbs down as a reaction to help us prioritize this feature. Feel free to comment if you have a particular need or comment on how this feature should work.
The text was updated successfully, but these errors were encountered:
An alternative would be to provide an alternative health check such as /healthz which only returns a 200. This way we can get a very quick response and save /api.status for when more detailed information is needed about the overall operational status, especially since the cost of running /api.status is not free.
Enhance /api/status to return multiple status codes or more information
Problem
Today the status endpoint returns
200
or500
and if you use this endpoint as a health check for a LB or other node manager, the node may get killed prematurely.A
500
generally indicates one of our metrics has exceeded a prescribed threshold, most commonly a db connection wait, or similar issue.This can simply mean the box is under heavy load, and not necessarily that the box is sick.
Solution
Either provide some configuration as to what the status endpoint returns or reserve
500
for a more catastrophic error - and provide some varying degrees of status so it functions better as a health check for a LB or other node manager.Alternatives/workarounds
If the
/api/status
endpoint is causing you trouble as a health check endpoint, you can also make an API call such as/api/login
with a known user + password w/out anapplicationId
- as a health check.Additional context
N/A
Related
#362
How to vote
Please give us a thumbs up or thumbs down as a reaction to help us prioritize this feature. Feel free to comment if you have a particular need or comment on how this feature should work.
The text was updated successfully, but these errors were encountered: