Set up a metrics dashboard #222
Conversation
|
What the dashboard looks like:
|
arunasank
changed the title
lib/dashboard.js
Outdated
| properties: { | ||
| view: 'timeSeries', | ||
| stacked: false, | ||
| title: 'WatchbotQueue: ApproximateNumberOfMessagesNotVisible, ApproximateNumberOfMessagesVisible', |
There was a problem hiding this comment.
Nitpick: could we make this name shorter? I don't think the full title would fit in the dashboard UI. Could we maybe change it to Visible and NotVisible Messages?
lib/dashboard.js
Outdated
| stacked: false, | ||
| title: 'WatchbotQueue: NumberOfMessagesDeleted', | ||
| metrics: [ | ||
| ['AWS/SQS', 'NumberOfMessagesDeleted', 'QueueName', '${AWS::StackName}-WatchbotQueue', { period: 60 }] |
There was a problem hiding this comment.
Is there a place to specify the Statistic to assess for this metric? I believe we want the Sum here, and the screenshot you made was using the Average.
lib/dashboard.js
Outdated
| metrics: [ | ||
| ['Mapbox/ecs-cluster', 'RunningCapacity', 'ClusterName', '${Cluster}', 'ServiceName', '${WatchbotService}', { period: 60 }], | ||
| ['.', 'DesiredCapacity', '.', '.', '.', '.', { period: 60 }], | ||
| ['AWS/SQS', 'NumberOfMessagesDeleted', 'QueueName', '${AWS::StackName}-WatchbotQueue', { period: 60 }] |
There was a problem hiding this comment.
Again, we'll want this metric to be Sum.
lib/dashboard.js
Outdated
| properties: { | ||
| view: 'timeSeries', | ||
| stacked: false, | ||
| title: 'Scaling as a function of the rate of processing', |
There was a problem hiding this comment.
Since scaling isn't triggered by the rate of processing, I think the use of as a function of here is misleading. Could we change the name of this title to something like Concurrency vs. Throughput, or possibly Capacity vs. Progress
Taking a step back, do we need this graph? We have the capacity on the graph above, and the NumberOfMessagesDeleted to the left of this graph. I feel like this graph may confuse people.
There was a problem hiding this comment.
imho this graph is the most useful, since this is the one most of us used for understanding scaling behaviour, and I feel pretty strongly about keeping it for that reason to spare folks the effort and time of generating it/figuring out what metrics are most relevant to generate this.
👍 on the name change and the statistic comments. I need to do a final check of the names and statistics - will fix these.
|
@jakepruitt I also want to add some DeadLetterQueue metrics, specifically the |
Fixes #216
cc/ @mapbox/platform-engine-room
TODO