Add ability to snooze alerts per universe for a specified duration #2891
Assignees
Labels
area/platform
Yugabyte Platform
Comments
iSignal
added a commit
that referenced
this issue
Jan 27, 2020
Summary: Multiple improvements requested by customers for 2.1. 1/ API to disable alert emails per universe for a given time interval or permanently. 2/ Send out alerts email in JSON format 3/ App config to report only errors in alert emails For (1) the API request looks like ``` curl --data 'disabled=true&disablePeriodSecs=3600' -X POST -H'X-AUTH-TOKEN: XXXXX' http://localhost:9000/api/v1/customers/f33e3c9b-75ab-4c30-80ad-cba85646ea39/universes/dadbd2bc-c202-4e8d-bf9d-a0a214ea3c4f/config_alerts ``` disablePeriodSecs is an optional parameter. If disabled=false, this parameter is ignored even if set. For (2), the current email is now sent as a multipart MIME message. The plaintext part has the JSON email so users can retrieve and parse it if needed. For (3), the new true/false config setting is a check box in the customer profile (customerconfig). The behavior of alert emails is now as follows: Emails are sent in the following cases: * Every status update period, a full list of nodes and checks is sent. * Every health check interval, an email is sent only if errors are discovered. The email has the full list of nodes and checks unless report_only_errors is set in the profile, in which case, only errors are reported. ** The first success after errors clear also results in an email. The email has the full list of nodes and checks unless report_only_errors is set in the profile, in which case, the email body just says "No errors to report". Test Plan: Multiple test cases For (1), there is a unit test that covers the key scenarios. I also verified manually that the correct arguments are passed to the python script. For (2), I manually inspected the multipart MIME message and verified it had the right JSON text. For (3), I simulated errors on the tserver in the python script and verified that only the errors were sent out with the setting enabled. I also verified that the success email gets sent out when errors stop. Reviewers: arnav, ram, bogdan Reviewed By: ram Subscribers: jenkins-bot, yugaware Differential Revision: https://phabricator.dev.yugabyte.com/D7816
iSignal
added a commit
that referenced
this issue
Jan 27, 2020
…eck panel Summary: https://phabricator.dev.yugabyte.com/D7816 has the API changes. This diff adds a way to show a message in the UI if alerts are disabled. Test Plan: Set alerts to enabled, disabled permanently and disabled until 1 hour. Verify that the message shows as expected Reviewers: ram, arnav, andrew Reviewed By: andrew Subscribers: yugaware Differential Revision: https://phabricator.dev.yugabyte.com/D7838
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This would be helpful during operational issues or upgrades etc.
The text was updated successfully, but these errors were encountered: