feat(autofix): Duplicate check for automation runs

[Mihir-Mavalankar]

PR Details

• For the V0 launch of triage signals we are changing the conditions for starting an autofix run.
• More details explanation here: https://www.notion.so/sentry/Triage-Signals-V0-Technical-Implementation-Details-2a18b10e4b5d8086a7ceddaf4194849a?source=copy_link#2a18b10e4b5d80d2ab2ce33d12db52e0
• As mentioned in the doc we can't rely on seer_fixability_score alone going forward and hence this PR adds cache based check + seer_autofix_last_triggered as a check.
• A new column is overkill for this. @wedamija was right to point it out in this PR. Adding a redis cache based check is enough.

[JoshFerge]

could we add a test?

[kddubey]

the ttl should prolly be minutes, not 1 hour. the automation task has a 35 sec. limit w/ 1 retry. if something flakily fails in the automation task (e.g., gemini 429s) and we don't actually trigger autofix, we'll be delaying the next automation run for 1 hour. during that time it's possible someone would've wanted an RCA on their issue

if the ttl is minutes, then as long as a new event comes in after that, autofix has another chance to get triggered

or, if the automation task fails, delete the key. that feels a little riskier tho

[Mihir-Mavalankar]

the ttl should prolly be minutes, not 1 hour. the automation task has a 35 sec. limit w/ 1 retry. if something flakily fails in the automation task (e.g., gemini 429s) and we don't actually trigger autofix, we'll be delaying the next automation run for 1 hour. during that time it's possible someone would've wanted an RCA on their issue

if the ttl is minutes, then as long as a new event comes in after that, autofix has another chance to get triggered

or, if the automation task fails, delete the key. that feels a little riskier tho

Okay will change to 15 minutes

[kddubey]

nit: could prolly be 5 minutes

[Mihir-Mavalankar]

nit: could prolly be 5 minutes

Hmm 5 seems a bit too close. Will take it down to 10.