LAMINAR
New Migration
operations

Production PostgreSQL Migration

Running

prod-db-east → prod-db-west · PostgreSQL 15 · 214 datasets · running for 2h 29m

Migration progress
67.4%997.5 GB of 1.45 TB
ETA 9m at current rate
Consistency: Snapshot + logical replication tail
Current throughput
845 MB/s
Est. safe capacity
960 MB/s
Controller confidence
91%
model converged
Concurrency
28 / 64
workers × shards
p99 latency
8.4 ms
baseline 8.4 ms
Throttling
None
0 events this run
Adaptive Control Loop

One loop, running continuously against every target.

Stage · Observe
  1. 01
    Observe
  2. 02
    Estimate
  3. 03
    Probe
  4. 04
    Learn
  5. 05
    Adapt
  6. 06
    Checkpoint
  7. 07
    Recover
Rate control

Throughput, capacity ceiling and latency

Amber markers are rate changes issued by the controller. Utilisation of the current estimate is 88.0%.

Actual throughputEstimated safe capacityTarget ratep99 latencyRate change
Controller Decisions

Every rate change, with the observation that caused it.

  • 22:25:08INCREASE+8%

    Increased throughput by 8% — target latency remained stable across two probe windows.

  • 22:26:03HOLD

    Held throughput — feedback indicated rising latency at the current rate.

  • 22:27:07REDUCE-15%

    Reduced throughput by 15% after throttling was detected on the destination.

  • 22:27:28CHECKPOINT

    Checkpoint written — partition-8421 verified.

  • 22:27:47PROBE+3%

    Probe window opened — nudging 3% above estimate to keep the capacity model honest.

Recovery
Last verified checkpoint
partition-8421
995.1 GB verified
22:25:55 UTC · 3 minutes ago

On interruption Laminar resumes from the last verified checkpoint. Worst case is minutes of replay, not a restart.

  • partition-8421995.1 GB22:25:55
Audit timeline
  • 19:59:55Operation started

    prod-db-east → prod-db-west, adaptive mode: Balanced (BBR-style probing)

  • 20:00:35Capacity model initialised

    Initial estimate 960 MB/s from 40s observation window