Services:

One EC2 server (Start with the free one and see how it goes) serving pages and some mappings.
General Purpose SSD (gp3) holding recently accessed crash logs and some mappings.
Tiered S3 holding old crash logs.

Operation:

EC2 server receives a Crashy request.
- The requested ID is checked against the local SSD. If it doesn't exist, the 'no such crash log' page is returned.
- If the crash does exist, the React app is returned, which is templated to accept the crash data.
- If the crash was gotten from the S3 it's put back into the SSD.
- The SSD maintains a bidirectional map of 'Crash Id' to `Last Access Day'. This means that when a crash is returned this map is updated.
  - Every day or at will, all crash logs with last access day of <CurrentDay - 30> will be transferred to the S3 where it will be tiered automatically by Amazon.
In some cases the EC2 server will receive a mappings data request. The server will request the mappings from the appropriate provider and cache it for 30 days in the SSD. (Works the same way as crash logs, a bidirectional map is needed as well)
EC2 server supports upload crash and delete crash endpoints.