[db] Request database optimizations

[SeungjinYang]

The PR introduces a series of changes to make accesses to request DB more efficient.

General changes

1. The database is given an index on created_at. This is because several queries in requests.py have ORDER BY created_at statements, which can be accelerated by an index.
2. RequestTaskFilter now has a sort parameter that can toggle the inclusion of ORDER BY created_at statement. This allows callers that do not need the results to be sorted to benefit from not sorting the result.
3. RequestTaskFilter had fields parameter introduced 2 days ago. This PR adds appropriate fields parameter to callers that do not need all of the parameters (especially request / response bodies) to perform their tasks.
4. get_request_tasks_with_fields_async is merged with get_request_tasks_async, allowing the latter function to handle an optional fields parameter if provided. Similarly, get_request_tasks is modified to handle an optional fields parameter.
5. exact_match fields are added to some queries that acts on a request given a request_id. Since the API server wants to handle clients submitting a request ID prefix, query functions use WHERE request_id LIKE <prefix>% statement to handle prefixes. However, in cases where we know an exact request ID is supplied, using WHERE request_id = <id> is more efficient.

Case studies of specific codepaths

sky api cancel -a:

• Uses kill_requests. Since no request IDs are specified, the request IDs are retrieved from DB. This DB call now only returns request IDs (instead of whole requests) and does not sort. I expect this to be the bulk of the efficiency improvement.
• Since the request IDs are retrieved from DB, we can set exact_match to True on update_request. This uses an exact match query (WHERE request_id = <id> instead of WHERE request_id LIKE <prefix>%) making the operation more efficient.

sky logs

• uses _tail_log_file. We now establish an exact_request_id at the start of _tail_log_file, and use exact match query making the operation more efficient.

Tested (run the relevant ones):

• Code formatting: install pre-commit (auto-check on commit) or bash format.sh
• Any manual or new tests for this PR (please specify below)
• All smoke tests: /smoke-test (CI) or pytest tests/test_smoke.py (local)
• Relevant individual tests: /smoke-test -k test_name (CI) or pytest tests/test_smoke.py::test_name (local)
• Backward compatibility: /quicktest-core (CI) or pytest tests/smoke_tests/test_backward_compat.py (local)

[SeungjinYang]

/quicktest-core --base-branch v0.10.1
/smoke-test --kubernetes --no-resource-heavy

[SeungjinYang]

/smoke-test --aws --no-resource-heavy

[SeungjinYang]

/quicktest-core --base-branch v0.10.1
/smoke-test --kubernetes --no-resource-heavy
/smoke-test --aws --no-resource-heavy

[SeungjinYang]

There's a lot of stuff going on in this PR, so I'll separate the changes out and submit incremental PRs as recommended by @rohansonecha