Feat; Create new table for collector configuration #2157
Conversation
database/src/pool/postgres.rs
Outdated
| r#"CREATE EXTENSION IF NOT EXISTS "uuid-ossp";"#, | ||
| r#" | ||
| CREATE TABLE IF NOT EXISTS collector_config ( | ||
| id UUID PRIMARY KEY, |
There was a problem hiding this comment.
We will have probably max. 4? collectors in the near future 😅 So I think that UUIDs are overkill. My experience with messing with Postgres extensions is that it can go wrong, I would prefer to avoid this and just use good ol' autoincremented IDs.
On the other hand, I'd like to assign human-readable names to the collectors, for easier debugging. So please add a name TEXT NOT NULL column, thanks.
There was a problem hiding this comment.
Thanks! I think that the table looks good like it is, but also I don't think we necessarily need to merge this outright.
So I'd keep this lying here until we have some actual code (probably first in the website) that actually needs to use this table, and then write the corresponding DbPool queries so that we have a user of the table. Maybe we'll figure out that we need some additional columns by then or something.
database/schema.md
Outdated
| ``` | ||
| sqlite> SELECT * FROM collector_config; | ||
|
|
||
| id target date_added last_heartbeat_at benchmark_set is_active |
There was a problem hiding this comment.
Could you please update this with name? Thanks!
database/src/pool/sqlite.rs
Outdated
| @@ -404,6 +404,19 @@ static MIGRATIONS: &[Migration] = &[ | |||
| alter table pstat_series_with_target rename to pstat_series; | |||
| "#, | |||
| ), | |||
| Migration::without_foreign_key_constraints( | |||
There was a problem hiding this comment.
I wonder if we should "delay" the tables on the SQLite side until we finish the Postgres implementation? Just to potentially avoid needless migrations (i.e. Postgres could serve as a "testbed" for the DB). I'm not even sure if we'll ever need this table on the SQLite side, tbh.
UUIDextension for postgresactiveand able to take work. Has a heartbeat column so we can see if the collector goes offline.benchmark_setis for when we start splitting work between collectors