Postgresql plugin does not support SCRAM-SHA-256 authentication

[vitalca]

Relevant telegraf.conf:

[[inputs.postgresql]]
    address = "postgres://$PG_USERNAME:$PG_PASSWORD@$PG_HOST/$PG_DBNAME"
    databases = ["$PG_DBNAME"]

System info:

• telegraf 1.8 (official Docker image)
• Postgres 12

Steps to reproduce:

1. enable SCRAM-SHA-256 authentication

   ALTER SYSTEM SET password_encryption = 'scram-sha-256';

2. update passwords to recalculate hashes

   ALTER ROLE username PASSWORD 'xxx';

3. update pg_hba.conf
4. reload configs

   SELECT pg_reload_conf();

5. try to start telegraf

Expected behavior:

No issues

Actual behavior:

E! Error in plugin [inputs.postgresql]: unknown authentication type: 10

Temporary workaround:

Configure trust authentication for telegraf in the pg_hba.conf

[danielnelson]

This would be a good addition, do you think it is something you could work on?

related: https://github.com/jackc/pgx/commit/5044e8473ad948114b6cb63f6f30f94fc7834667

[powersj]

Hi,

It looks like this was added to the library Telegraf uses here. I tried launching a container with docker run --rm --name postgres --net host -e POSTGRES_USER=postgres -e POSTGRES_PASSWORD=password -e POSTGRES_INITDB_ARGS=--auth-host=scram-sha-256 postgres and was able to connect successfully:

postgresql,db=postgres,host=ryzen,server=dbname\=postgres\ host\=localhost\ user\=postgres tup_inserted=0i,datname="postgres",sessions_fatal=0i,sessions_abandoned=0i,tup_updated=0i,xact_commit=5i,blk_read_time=0,datid=13757i,deadlocks=0i,blks_read=171i,active_time=0.414,temp_bytes=0i,tup_returned=2907i,numbackends=1i,tup_deleted=0i,blks_hit=3038i,xact_rollback=0i,blk_write_time=0,temp_files=0i,sessions_killed=0i,sessions=2i,session_time=5.272,tup_fetched=1520i,conflicts=0i,idle_in_transaction_time=0 1660678797000000000
postgresql,db=postgres,host=ryzen,server=dbname\=postgres\ host\=localhost\ user\=postgres buffers_alloc=73i,buffers_backend_fsync=0i,checkpoints_req=1i,buffers_clean=0i,checkpoint_write_time=1,buffers_checkpoint=0i,maxwritten_clean=0i,buffers_backend=0i,checkpoint_sync_time=2,checkpoints_timed=0i 1660678797000000000

I'm going to mark this as closed, but if you are still seeing this issue, please do reply and let us know along with how to reproduce and telegraf logs.

Thanks!