Reference
Limits
Every fixed limit in Galena, in one place.
Content
| What | Limit |
|---|---|
| Component and group names | 100 characters |
| Component description | 500 characters |
| Monitor name | 100 characters |
| Monitor keyword | 200 characters |
| Accepted status codes per monitor | 20 |
| Incident and window titles | 200 characters |
| Update and window messages | 5,000 characters |
| Components per incident, window, subscriber or endpoint | 100 |
| Endpoint name | 100 characters |
| Endpoint URL | 2,000 characters, https:// only |
Checks
| What | Limit |
|---|---|
| Check interval | 1 minute |
| Timeout | 1 to 10 seconds, default 10 |
| Redirects followed | 5 |
| Body read for a keyword | the first 1 MiB |
| Checks run at once in one probe run | 50 |
| Probe run time | 50 seconds |
| Check results kept | 90 days |
| Monitors, with the default DynamoDB capacity | about 50 on dev, about 200 on prod, across three regions |
A probe run that waits on many slow targets at once can run out of its 50 seconds: 50 checks at a time, each allowed up to 10 seconds. Above roughly 200 monitors that becomes a real risk during a wide outage. Splitting a region's monitors across several runs is not built yet.
Rates
| What | Limit |
|---|---|
| API requests | 50 a second, bursts of 100 |
| Sign-in attempts | rate limited per visitor address |
| Subscribe form | 10 sign-ups per visitor address an hour; one confirmation email per address every 10 minutes; 200 confirmation emails an hour in total |
| Webhook and Slack sends | 8-second timeout; 5 attempts |
| Email sends | 5 attempts |
Past a subscribe limit, the form answers exactly as if it had worked, so it never reveals whether an address is subscribed.
History
| What | Kept |
|---|---|
| Days on the strip and in the uptime figure | 90 |
| Past incidents on the page and in the feeds | resolved in the last 14 days |
| Confirmation links | 7 days |
| Unsubscribe links | for good |