Closed
Bug 1141993
Opened 9 years ago
Closed 9 years ago
Pick & set thresholds for alerts using the rabbitmq New Relic plugin
Categories
(Tree Management :: Treeherder: Infrastructure, defect, P2)
Tree Management
Treeherder: Infrastructure
Tracking
(Not tracked)
RESOLVED
FIXED
People
(Reporter: emorley, Assigned: emorley)
References
Details
Now that bug 1093757 has installed the rabbitmq plugin, we can set up alerts for a number of rabbitmq related things: * Messages Available (# messages) * Publish Rate (messages/sec) * Delivery Rate (messages/sec) * Open Channels * Consumers We'll need to leave the plugin running for a while to figure out sensible thresholds for these for both prod and stage. https://rpm.newrelic.com/accounts/677903/dashboard/6293241/page/4
Assignee | ||
Comment 1•9 years ago
|
||
Done today as part of dealing with bug 1152681. Prod + stage: * Messages Available (# messages): warn=50, crit=100 * Publish Rate (messages/sec): warn=60, crit=100 * Delivery Rate (messages/sec): warn=60, crit=100 * Open Channels: warn=80 * Consumers: warn=60 The "Messages Available" stat is the most critical. To change them later, use the settings cog on: https://rpm.newrelic.com/accounts/677903/plugins/16138 The email alerts for plugins don't use the new alert section, so have to be set up here: https://rpm.newrelic.com/accounts/677903/integrations?page=alerting#tab-integrations=_mobile_amp_plugin_alert_email I've added my email to start with, and once I know how noisy it will/won't be (eg the text help on that page contradicts itself a few times - not clear what other non-plugin notifications will be sent to the email there as well) I can try adding others.
Assignee: nobody → emorley
Blocks: 1152681
Status: NEW → RESOLVED
Closed: 9 years ago
Priority: P3 → P2
Resolution: --- → FIXED
Assignee | ||
Comment 2•9 years ago
|
||
But the alerts/thresholds themselves seem to be working well, eg see the red shaded areas here: https://rpm.newrelic.com/accounts/677903/dashboard/6293241?tw[dur]=last_6_hours&tw[end]=1428587506
Assignee | ||
Comment 3•9 years ago
|
||
We were getting alerts on stage, since the recent new queue additions have bumped up the totals. As such I've updated stage/prod to the following: Note: higher values=worse (ie: the alert goes off when the actual value goes above the alert value). (In reply to Ed Morley [:emorley] from comment #1) > * Open Channels: warn=80 warn=110, crit=130 > * Consumers: warn=60 warn=65, crit=75 These values had previously been updated to be more strict, including here for reference: > * Publish Rate (messages/sec): warn=60, crit=100 warn=50, crit=80 > * Delivery Rate (messages/sec): warn=60, crit=100 warn=50, crit=80
You need to log in
before you can comment on or make changes to this bug.
Description
•