Investigate max_wal_size configuration in Railway

[lightwalker-eth]

I noticed the above being flagged in the logs.

Additionally, note the messages about: "checkpoints are occurring too frequently (18 seconds apart)"

[BanaSeba]

@lightwalker-eth Those are not errors, it's a regular Postgres procedure (Railway seems to not understand this as this is using LOG ?).
Time between those is not dramatic considering how much I/O stuff is going on during the indexing process. I would not suggest adjusting max_wal_size. This should not have a bigger impact on the indexing process.
I would consider this task as ready to close.

[lightwalker-eth]

@BanaSeba Thanks, appreciate your advice and understand these aren't errors, but rather interpret them as warnings of a potential optimization opportunity.

Could you please investigate what max_wal_size is actually being applied in our Railway deployment of Postgres?

[BanaSeba]

@lightwalker-eth max_wal_size is using the default value (1GB). Referring to the official doc This parameter can only be set in the postgresql.conf file or on the server command line.

• Tuning this will not have a meaningful impact on indexing
• It takes bigger work to set this in Railway (probably using our predefined Docker image)