feat: add timeout for query

[jiacai2050]

Which issue does this PR close?

Closes #165

Rationale for this change

See issue above

This PR only support set timeout in config file, will support set timeout in request level in next PR.

What changes are included in this PR?

In principle, if all operations are async/await, we can just timeout at the beginning of query, but it's not that ideal in real case, there maybe exists

• spawned background tasks, similar leak happens in Clean up spawned task on drop for AnalyzeExec, CoalescePartitionsExec, HashAggregateExec datafusion#1121
• stream made of iterators, which is a sync operation and cannot be cancelled from outside.

So we have to add timeout detect in those cases beside top-level timeout. AKAIK, there are two places need special handling:

• Memtable scan, include sync operation
• SST scan, include spawned tasks

When implement this feature, I also find TableProvider is wrongly cached, please see comments in

• https://github.com/CeresDB/ceresdb/pull/539/files#r1063260688
• https://github.com/CeresDB/ceresdb/pull/539/files#r1063262263

Are there any user-facing changes?

No

How does this change test

Existing UT, and manually test a large table with select count(1) from cpu;, it will return following error after 5s timeout

$ time curl --location --request POST 'http://0:5446/sql' --data-raw 'select count(1) from cpu;'

{"code":500,"message":"Failed to handle request, err:Query timeout, query:select count(1) from cpu;, err:deadline has elapsed"}
real    0m5.013s
user    0m0.008s
sys     0m0.000s

[ShiKaiWi]

LGTM