No description provided.
Bug 1436478 Comment 0 Edit History
Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.
Why? * A shared base of indexed data can allow services to make smarter decisions and save money on provisioning and data storage * Azure is unreliable with frequent failures and inscrutible rate-limiting. While we are resilient to these errors, they do cause inefficiency and sometimes cascading failures * Storing data in postgres will enable features users desire: * Queue introspection (what tasks are pending right now? how many tasks are [actually pending and not cancelled](https://bugzilla.mozilla.org/show_bug.cgi?id=1434851)?) * Integration of task-related and worker-related status (what tasks has this worker performed?) * Ad-hoc analytics (how much compute time do we use per push? what do we pay for wpt?)
Why? * A shared base of indexed data can allow services to make smarter decisions and save money on provisioning and data storage * Azure is unreliable with frequent failures and inscrutible rate-limiting. While we are resilient to these errors, they do cause inefficiency and sometimes cascading failures * Storing data in postgres will enable features users desire: * Queue introspection (what tasks are pending right now? how many tasks are [actually pending and not cancelled](https://bugzilla.mozilla.org/show_bug.cgi?id=1434851)?) * Integration of task-related and worker-related status (what tasks has this worker performed?) * Ad-hoc analytics (how much compute time do we use per push? what do we pay for wpt?) * More sophisticated dependency resolution [in the case of task failure](https://bugzilla.mozilla.org/show_bug.cgi?id=1443503#c8)
Why? * A shared base of indexed data can allow services to make smarter decisions and save money on provisioning and data storage * Azure is unreliable with frequent failures and inscrutible rate-limiting. While we are resilient to these errors, they do cause inefficiency and sometimes cascading failures * Storing data in postgres will enable features users desire: * Queue introspection (what tasks are pending right now? how many tasks are [actually pending and not cancelled](https://bugzilla.mozilla.org/show_bug.cgi?id=1434851)?) * Integration of task-related and worker-related status (what tasks has this worker performed?) * Ad-hoc analytics (how much compute time do we use per push? what do we pay for wpt?) * More sophisticated dependency resolution [in the case of task failure](https://bugzilla.mozilla.org/show_bug.cgi?id=1443503#c8) * Dynamically changing task levels