docs: update flags with additions, removals, and default changes

[vyavdoshenko]

• Add new flags: --logbuflevel, --logbufsecs, --cluster_coordinator_connect_timeout_ms, --cluster_coordinator_response_timeout_ms, --container_iteration_yield_interval_usec, --enable_memcache_io_loop_v2, --enable_resp_io_loop_v2, --journal_omit_redundant_writes, --listpack_max_bytes, --listpack_max_field_len, --s3_use_helio_client, --serialization_tagged_chunks, --tiered_max_pending_stash_bytes, --use_oah_set
• Remove deprecated/replaced flags: --tiered_storage_write_depth, --logtostdout, --experimental_io_loop_v2
• Update default values for --max_log_size (1800→200), --rdb_sbf_chunked (false→true)

[vercel]

@vyavdoshenko is attempting to deploy a commit to the DragonflyDB Team on Vercel.

A member of the Team first needs to authorize it.

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (2) 📘 Rule violations (0)

1. Stale logtostdout guidance 🐞 Bug ≡ Correctness

Description

docs/managing-dragonfly/cluster-mode.md still recommends passing --logtostdout, but this PR
removes that flag from the canonical flags reference, so readers will try to use an option that’s no
longer documented/supported.

Code

docs/managing-dragonfly/flags.md[L514-517]

-### `--logtostdout`
-  Log messages go to stdout instead of logfiles.
+### `--logbuflevel`
+  Buffer log messages logged at this level or below. (-1 means don't buffer; 0 means buffer INFO only).

-  `default: false`
+  `default: 0`
+
+### `--logbufsecs`
+  Buffer log messages for at most this many seconds.
+
+  `default: 30`

Evidence

The cluster-mode guide explicitly instructs --logtostdout, while the updated flags reference now
documents other logging flags (--logtostderr, --alsologtostderr, --logbuf*) and no longer
includes --logtostdout.

docs/managing-dragonfly/cluster-mode.md[195-202]
docs/managing-dragonfly/flags.md[494-517]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

### Issue description
`docs/managing-dragonfly/cluster-mode.md` tells users to pass `--logtostdout` to debug `DFLYCLUSTER CONFIG` failures, but the flag is removed from the flags reference in this PR. This creates stale guidance that will confuse users.

### Issue Context
The flags reference already documents `--logtostderr` / `--alsologtostderr` as alternatives.

### Fix Focus Areas
- docs/managing-dragonfly/cluster-mode.md[195-202]
- docs/managing-dragonfly/flags.md[494-517]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

2. Tiering doc references removed flag 🐞 Bug ≡ Correctness

Description

docs/managing-dragonfly/tiering.md still lists tiered_storage_write_depth, but this PR removes
that flag from the synchronized flags reference, leaving the tiering guide out of date.

Code

docs/managing-dragonfly/flags.md[L289-293]

-### `--tiered_storage_write_depth`
-  Maximum number of concurrent stash requests issued by background offload. 
-
-  `default: 200`
-

Evidence

The tiering guide still instructs tiered_storage_write_depth, but the current flags reference no
longer lists it among the other tiered_* options; additionally, the docsync documentation explains
that flags are deleted from flags.md when the server no longer reports them, so the reference
removal implies obsolescence.

docs/managing-dragonfly/tiering.md[32-44]
docs/managing-dragonfly/flags.md[273-293]
tools/docsync/README.md[89-107]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

### Issue description
`docs/managing-dragonfly/tiering.md` documents `tiered_storage_write_depth` as a main tiering flag, but the canonical `flags.md` reference no longer contains this flag. This makes the tiering guide’s “main flags” list inconsistent with the current flag surface.

### Issue Context
This repo’s docsync tooling treats the running server as authoritative and deletes flags from `flags.md` when the server no longer reports them.

### Fix Focus Areas
- docs/managing-dragonfly/tiering.md[32-44]
- docs/managing-dragonfly/flags.md[273-293]
- tools/docsync/README.md[89-107]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

[qodo-free-for-open-source-projects]

PR Summary by Qodo

Docs: refresh CLI flags list (additions, removals, and default updates)
📝 Documentation 🕐 Less than 10 minutes

Walkthroughs

Description

• Document newly introduced CLI flags across logging, clustering, IO loops, and tiered storage.
• Remove entries for deprecated/replaced flags to match current CLI surface.
• Update documented defaults for --max_log_size and --rdb_sbf_chunked.

High-Level Assessment

The following are alternative approaches to this PR:

1. Auto-generate flags.md from the flag registry / `--help` output

• ➕ Prevents drift between documented flags and actual binary flags/defaults
• ➕ Makes adding/removing flags update docs automatically in CI
• ➖ Requires tooling and a stable machine-readable flag schema/output
• ➖ Generated docs may be harder to keep human-curated/annotated

2. Add a CI check that diffs `--help` against docs

• ➕ Lower effort than full generation
• ➕ Catches missing/incorrect documentation early
• ➖ Still leaves manual editing as the primary workflow
• ➖ Comparing help output to markdown can be noisy/fragile

Recommendation: For this PR, the manual docs sync is appropriate and low-cost. Longer-term, consider either auto-generating the flags table (best drift-prevention) or adding a CI drift check as a lighter-weight safeguard.

File Changes

Documentation (1)

flags.md Sync documented CLI flags with current additions, removals, and defaults +70/-15

Sync documented CLI flags with current additions, removals, and defaults

• Adds documentation entries for multiple new flags (logging buffering, coordinator timeouts, IO loop toggles, serialization/tiering controls, and listpack limits). Removes deprecated/replaced flags ('--tiered_storage_write_depth', '--logtostdout', '--experimental_io_loop_v2'). Updates documented defaults for '--max_log_size' (1800 → 200) and '--rdb_sbf_chunked' (false → true).

docs/managing-dragonfly/flags.md

---