I'm currently working on monitoring PostgreSQL logs in CSV format using mtail. One challenge is that SQL statements in the statement field may span multiple lines. This makes it difficult to extract meaningful labels or metrics using mtail, since it appears to process logs line-by-line.

For example, a single statement field in a PostgreSQL log might contain embedded newlines like this:

2025-06-26 01:47:40.578 CST,,,2348220,,685c3414.23d4bc,1,,2025-06-26 01:38:28 CST,509/2455864,0,LOG,00000,"automatic vacuum of table ""bdc.pg_catalog.pg_attribute"": index scans: 1
pages: 0 removed, 2420645 remain, 0 skipped due to pins, 2391644 skipped frozen
tuples: 18257 removed, 1605338 remain, 1419055 are dead but not yet removable, oldest xmin: 3862849561
buffer usage: 1441377 hits, 1443437 misses, 2468 dirtied
avg read rate: 22.069 MB/s, avg write rate: 0.038 MB/s
system usage: CPU: user: 9.42 s, system: 8.80 s, elapsed: 510.99 s",,,,,,,,,""

I’m wondering:

• Does mtail support any way to accumulate or buffer multiple lines to allow multi-line matching?
• If not directly supported, are there recommended workarounds for this scenario?

This would be very helpful when we want to extract meaningful metrics or tags (e.g., command type, table name) from full SQL statements for observability purposes.

Thanks in advance for your guidance!