[DO NOT MERGE] updated: storage benchmarks

[aramikm]

Goal

The goal of this PR is to update the database benchmarks

Closes #2606

Discussion

• The read values are substantially increased but it appears that the POV is still the bottle-neck for the most part for weights and not the ref time.
• I had to update the stable weights since they were going to change regardless.
• Here is the calculations solely based on ref time

Some considerations

• Providers might ran out of capacity faster due to increase in weights

Checklist

• Unit Tests updated
• Spec version incremented?

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[JoeCap08055]

Also to note, it's not just the increase in MSAs that contributes to the growth in keys. Freesky is posting tons of content; right now on testnet there are over 3.4M message keys and growing daily at a significant clip. This is significant, because while the growth in new accounts and graphs may plateau at points, presumably the amount of content will continue to grow unbounded...

[shannonwells]

Can we estimate, based on the freesky usage, how comparatively fast capacity will drain based on these new weights? Might be good to have numbers like that somewhere. Not a blocking comment obviously.

[aramikm]

Tried multiple attempts to fix this test but for it always found a way to fail in one way or another. Would appreciate if you know how to fix this.

[aramikm]

Moving all of them into sudo.test.ts because if a test is using sudo key it should be inside that file so that it doesn't error in parallel mode.

[aramikm]

Can we estimate, based on the freesky usage, how comparatively fast capacity will drain based on these new weights? Might be good to have numbers like that somewhere. Not a blocking comment obviously.

Based on my calculations presented in the PR add_ipfs_message is increased around ~7x so I would assume that they will need around 7x of their current max used capacity

[aramikm]

Pallet base reads

2025-09-26 20:13:26  creating SingleState txpool Limit { count: 8192, total_bytes: 20971520 }/Limit { count: 512, total_bytes: 1048576 }.
2025-09-26 21:13:22 Warmup round 1/2
2025-09-27 00:02:04 Warmup round 2/2
2025-09-27 02:25:04 Preparing keys from block 0x7abe…8bd7
2025-09-27 02:25:04 starting metrics for msa
2025-09-27 02:29:35 Reading 6346291 keys
2025-09-27 02:36:54 Pallet Time summary [ns]:
Total: 434471377632
Min: 3770, Max: 31897572
Average: 68460, Median: 58485, Stddev: 82166.91
Percentiles 99th, 95th, 75th: 242181, 87478, 65701
Value size summary:
Total: 95733143
Min: 1, Max: 138
Average: 15, Median: 8, Stddev: 15.36
Percentiles 99th, 95th, 75th: 35, 35, 35
2025-09-27 02:36:54 starting metrics for schemas
2025-09-27 02:36:54 Reading 70 keys
2025-09-27 02:36:54 Pallet Time summary [ns]:
Total: 1200539
Min: 6790, Max: 62902
Average: 17150, Median: 15230, Stddev: 7620.46
Percentiles 99th, 95th, 75th: 62902, 27710, 19871
Value size summary:
Total: 12930
Min: 2, Max: 1367
Average: 184, Median: 5, Stddev: 286.21
Percentiles 99th, 95th, 75th: 1367, 700, 414
2025-09-27 02:36:54 starting metrics for messages
2025-09-27 02:37:09 Reading 1605421 keys
2025-09-27 02:38:20 Pallet Time summary [ns]:
Total: 70052453862
Min: 2109, Max: 1809357
Average: 43634, Median: 42268, Stddev: 17860.76
Percentiles 99th, 95th, 75th: 68231, 58795, 48476
Value size summary:
Total: 81876373
Min: 2, Max: 51
Average: 50, Median: 51, Stddev: 0.05
Percentiles 99th, 95th, 75th: 51, 51, 51
2025-09-27 02:38:20 starting metrics for handles
2025-09-27 02:42:49 Reading 6013797 keys
2025-09-27 02:50:27 Pallet Time summary [ns]:
Total: 453641114292
Min: 16910, Max: 37473914
Average: 75433, Median: 55114, Stddev: 134426.98
Percentiles 99th, 95th, 75th: 867059, 83356, 61523
Value size summary:
Total: 61214021
Min: 4, Max: 34
Average: 10, Median: 8, Stddev: 6.13
Percentiles 99th, 95th, 75th: 25, 22, 15
2025-09-27 02:50:27 starting metrics for capacity
2025-09-27 02:50:27 Reading 50 keys
2025-09-27 02:50:27 Pallet Time summary [ns]:
Total: 18421944
Min: 1395, Max: 15846482
Average: 368438, Median: 13308, Stddev: 2218895.41
Percentiles 99th, 95th, 75th: 15846482, 913746, 18029
Value size summary:
Total: 1602
Min: 1, Max: 101
Average: 32, Median: 32, Stddev: 23.32
Percentiles 99th, 95th, 75th: 101, 101, 41
2025-09-27 02:50:30 Time summary [ns]:
Total: 958184568269
Min: 1395, Max: 37473914
Average: 68610, Median: 55534, Stddev: 104861.07
Percentiles 99th, 95th, 75th: 743669, 82813, 62747
Value size summary:
Total: 238838069
Min: 1, Max: 1367
Average: 17, Median: 8, Stddev: 16.69
Percentiles 99th, 95th, 75th: 51, 51, 35
2025-09-27 02:50:30 Writing weights to "/home/ubuntu/rocksdb_weights.rs"

[aramikm]

No child tree read weights

//! THIS FILE WAS AUTO-GENERATED USING THE SUBSTRATE BENCHMARK CLI VERSION 48.0.0
//! DATE: 2025-09-26 (Y/M/D)
//! HOSTNAME: `ip-10-173-4-131`, CPU: `Intel(R) Xeon(R) Platinum 8488C`
//!
//! DATABASE: `RocksDb`, RUNTIME: `Frequency`
//! BLOCK-NUM: `BlockId::Number(9443452)`
//! SKIP-WRITE: `true`, SKIP-READ: `false`, WARMUPS: `2`
//! STATE-VERSION: `V1`, STATE-CACHE-SIZE: ``
//! WEIGHT-PATH: ``
//! METRIC: `Average`, WEIGHT-MUL: `1.0`, WEIGHT-ADD: `0`

// Executed Command:
//   ./frequency/target/release/frequency
//   benchmark
//   storage
//   --state-version=1
//   --chain=frequency
//   --base-path=/data
//   --warmups=2
//   --skip-write

/// Storage DB weights for the `Frequency` runtime and `RocksDb`.
pub mod constants {
	use frame_support::weights::constants;
	use sp_core::parameter_types;
	use sp_weights::RuntimeDbWeight;

	parameter_types! {
		/// By default, Substrate uses `RocksDB`, so this will be the weight used throughout
		/// the runtime.
		pub const RocksDbWeight: RuntimeDbWeight = RuntimeDbWeight {
			/// Time to read one storage item.
			/// Calculated by multiplying the *Average* of all values with `1.0` and adding `0`.
			///
			/// Stats nanoseconds:
			///   Min, Max: 1_395, 37_473_914
			///   Average:  68_610
			///   Median:   55_534
			///   Std-Dev:  104861.07
			///
			/// Percentiles nanoseconds:
			///   99th: 743_669
			///   95th: 82_813
			///   75th: 62_747
			read: 68_610 * constants::WEIGHT_REF_TIME_PER_NANOS,

			/// Time to write one storage item.
			/// Calculated by multiplying the *Average* of all values with `1.0` and adding `0`.
			///
			/// Stats nanoseconds:
			///   Min, Max: ,
			///   Average:
			///   Median:
			///   Std-Dev:
			///
			/// Percentiles nanoseconds:
			///   99th:
			///   95th:
			///   75th:
			write: 0 * constants::WEIGHT_REF_TIME_PER_NANOS,
		};
	}

	#[cfg(test)]
	mod test_db_weights {
		use super::constants::RocksDbWeight as W;
		use sp_weights::constants;

		/// Checks that all weights exist and have sane values.
		// NOTE: If this test fails but you are sure that the generated values are fine,
		// you can delete it.
		#[test]
		fn bound() {
			// At least 1 µs.
			assert!(
				W::get().reads(1).ref_time() >= constants::WEIGHT_REF_TIME_PER_MICROS,
				"Read weight should be at least 1 µs."
			);
			assert!(
				W::get().writes(1).ref_time() >= constants::WEIGHT_REF_TIME_PER_MICROS,
				"Write weight should be at least 1 µs."
			);
			// At most 1 ms.
			assert!(
				W::get().reads(1).ref_time() <= constants::WEIGHT_REF_TIME_PER_MILLIS,
				"Read weight should be at most 1 ms."
			);
			assert!(
				W::get().writes(1).ref_time() <= constants::WEIGHT_REF_TIME_PER_MILLIS,
				"Write weight should be at most 1 ms."
			);
		}
	}
}