Helm chart: added configuration of uwsgi using environment variables

[alexintech]

What this PR does

Adds uwsgi configuration to helm chart.
Sets environment variables with name capitalized and prefixed with UWSGI_, and dashes are substituted with underscores like described here

Sets UWSGI_LISTEN=1024 by default, but can be overwritten or completely removed by uwsgi: null

Or is it better to not specify default value (it's not backward compatible)?

Also, small indentation fixes for postgresql configuration.

Which issue(s) this PR fixes

closes #562

Also, this PR has been closed because of this PR

Checklist

• Unit, integration, and e2e (if applicable) tests updated
• Documentation added (or pr:no public docs PR label added if not required)
• CHANGELOG.md updated (or pr:no changelog PR label added if not required)

[joeyorlando]

nice! thanks for your continued contribution to improving the Helm charts @alexintech

[jrossBah]

@alexintech Big question for ya. I am currently working with getting grafana oncall implemented and we had to turn this setting down to 128 to get it to work in our kubernetes cluster. I am wondering why you default set it to 1024. As We understand it the container the chart spins up defaults to 128. We found this by changing it on the ec2 instance and still seeing the error until we lowered this new uwsgi setting.

Do you know is there a way in these values that changes the container default? Thank you!

[alexintech]

@alexintech I am wondering why you default set it to 1024.

Do you know is there a way in these values that changes the container default? Thank you!

@jrossBah That's not my decision. 1024 has been hardcoded in helm chart. I only made this value configurable. To maintain the same behavior default value remained the same.

[jrossBah]

@alexintech I am wondering why you default set it to 1024.
Do you know is there a way in these values that changes the container default? Thank you!

@jrossBah That's not my decision. 1024 has been hardcoded in helm chart. I only made this value configurable. To maintain the same behavior default value remained the same.

That's so interesting. What we found is that this limit seems to conflict with the limitations of the kubernetes and also docker containers that set at 128. So unless you bring this down to 128 or under the engine pods will fail.