I feel that it should be merged, and the information in the Tag Wiki moved to query-performance or other postgres tags as appropriate:
Given that every single question that is tagged postgres-performance will also be tagged postgres, the only gain in having a separate postgres-performance tag is the ability to have a customized tag wiki.
The tag wiki itself is great, and especially in this middle paragraph, there is significant Postgres-specific advice:
Consider the basic advice on Performance Optimization and Slow Query Questions in PostgreSQL Wiki before asking questions on this tag, including the "Things to try before you post" section.
Using EXPLAINis particularly important. Useful ebooks are freely available.
The down-side is that it invites fragmentation of all other tags — do we really want to go down the line of having postgres-database-design, postgres-replication etc? The benefit would be the same for these cases too, namely that we could create tag wikis with Postgres-specific links and information.
Alternatively, why not just have a single 'performance' tag, and divide the tag wiki into sections for each RDBMS? All the text in the postgres-performance tag wiki could be transferred 'as is' into query-performance et al.
There are two bits of data we are lacking that might be helpful to have from Stack Exchange:
- How often is the tag wiki page visited.
- Which tagging system is more beneficial for search purposes.
No action needed
Benefits of the status quo
- The postgresql-performance tag is relatively popular (253 questions)
- It is actively curated and promoted by its subject area experts both here and on SO
- The tag Excerpt and Wiki are (unusually!) useful, appropriate, and targeted to the intended usage
This tag therefore adds value to the community in the practical way it is used.
Someone will need to review the result to:
- Ensure the new tag is actually appropriate in each case
- Add postgres to the question if needed
- Remove references (perhaps in comments) to the removed tag
This makes work for the subject area experts, without providing them with any compensating benefit.
It will also deprive them of a tool they explicitly created and actively use & maintain.
Why bother with this change?
The current tag appears useful to the people using it, and is not causing any actual harm to the site.
Making the change may inconvenience existing tag users, and require moderator time to implement.
The only benefit seems to be to eliminate a perceived technical redundancy in the tagging system.
It does not set a precedent. New tags are reviewed on their merits.
I think discussing / tuning / optimizing performance is sufficiently distinct for the various RDBMS to warrant a separate tag. We have put some thought into refining information, instructions and resources in postgresql-performance and it serves well so far. (I say "we" because it started with the related tag on SO.)
If general performance aspects are involved, it can make sense to additionally or instead tag one or more of performance, query-performance, performance-tuning - which also serve well for RDBMS that don't have a specific this-rdbms-performance (yet).
Other than that, I can subscribe to what Paul answered. The tag should stay.