Health check

[blackandred]

I think a more precised health checking should be created that would check the backend and frontend setup, the connectivity with SMTP and possibly with LDAP.

[zicklag]

That would be a good feature request for Taiga actually, to provide a /status page. I've seen that in other software before such as JIRA. Then healthcheck scripts only need to do a curl to /status to ensure the health of the application.

Not that we couldn't implement it ourselves because I have no idea how long it would be before they worked something like that in.

[blackandred]

I agree, each application should implement that. If the application does not implement an endpoint, then we can provide some basic checking.

…

On Tue, Aug 20, 2019 at 9:20 PM Zicklag ***@***.***> wrote: That would be a good feature request for Taiga actually, to provide a /status page. I've seen that in other software before such as JIRA. Then healthcheck scripts only need to do a curl to /status to ensure the health of the application. Not that we couldn't implement it ourselves because I have no idea how long it would be before they worked something like that in. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#6?email_source=notifications&email_token=AFOAJBRCNKQJALRB4QVKSHTQFQ7ZLA5CNFSM4INQA6JKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD4XMF7A#issuecomment-523158268>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFOAJBWCAODJS2JB5KRQ4PLQFQ7ZLANCNFSM4INQA6JA> .

[Leopere]

Even curl'ing an asset from the Taiga-Front like a favicon would be indicative of uptime. I don't know what you'd query for Taiga-Back but I feel a curl to the API would be where to begin there. Simply curling isn't perfect but it would at least show some degree of container health so if you want self-healing or dependencies you can have them rely on the run status of the two services.

[blackandred]

That's a good idea Leopere :)

…

On Sun, Aug 25, 2019 at 9:56 PM Leopere ***@***.***> wrote: Even curl'ing an asset from the Taiga-Front like a favicon would be indicative of uptime. I don't know what you'd query for Taiga-Back but I feel the API would be where to begin there. Simply curling isn't perfect but it would at least show some degree of container health so if you want self-healing or dependencies you can have them rely on the run status of the two services. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#6?email_source=notifications&email_token=AFOAJBQDRLB2NLWFDMMJ56TQGLPWNA5CNFSM4INQA6JKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD5C2RWA#issuecomment-524658904>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFOAJBXBTFTCL2PNKOIKZULQGLPWNANCNFSM4INQA6JA> .

[h44z]

in terms of container health checks i would recommend using https://github.com/aelsabbahy/goss.
this nice utility can check processes, api's and so on =)

then simply include a script:

#!/bin/bash

set -e

goss -g /app/goss.yaml validate --format json_oneline

exit 0

in the dockerfile:

HEALTHCHECK --interval=1m --timeout=10s \
    CMD /app/healthcheck.sh