Tô Gia Bảo — WritingTechnical writing on data systems, software design and applied AI.https://tgblab.io.vn/en-usStart With the Decision, Not the Stackhttps://tgblab.io.vn/writing/start-with-the-decision-not-the-stack/https://tgblab.io.vn/writing/start-with-the-decision-not-the-stack/A production data architecture starts with the decision it must support. Define the output grain, freshness, correctness and repair promise before choosing the tools.Wed, 09 Sep 2026 02:00:00 GMTData systemsCloud infrastructureSoftware designbaokdl2226@gmail.comProduction Must Confirm the Releasehttps://tgblab.io.vn/writing/production-must-confirm-the-release/https://tgblab.io.vn/writing/production-must-confirm-the-release/A rollout becomes the current release only when production evidence matches the intended artifact, configuration, schema and functional path.Thu, 13 Aug 2026 11:15:00 GMTCloud infrastructureSoftware designbaokdl2226@gmail.comDeployment Is a State Transitionhttps://tgblab.io.vn/writing/deployment-is-a-state-transition/https://tgblab.io.vn/writing/deployment-is-a-state-transition/A production change is safer when provisioning, schema compatibility, rollout and verification move through named states with explicit stop conditions.Thu, 13 Aug 2026 09:15:00 GMTCloud infrastructureSoftware designbaokdl2226@gmail.comA Container Needs a Runtime Contracthttps://tgblab.io.vn/writing/a-container-needs-a-runtime-contract/https://tgblab.io.vn/writing/a-container-needs-a-runtime-contract/A stable image becomes a reliable service only when configuration, secrets, state, probes and shutdown behaviour have explicit owners.Thu, 13 Aug 2026 07:15:00 GMTCloud infrastructureSoftware designbaokdl2226@gmail.comThe Image Is the Releasehttps://tgblab.io.vn/writing/the-image-is-the-release/https://tgblab.io.vn/writing/the-image-is-the-release/Build once, name the result by digest and let environments configure the same artifact instead of quietly creating different releases.Thu, 13 Aug 2026 04:15:00 GMTCloud infrastructureSoftware designbaokdl2226@gmail.comProving a Pipeline Is Replayablehttps://tgblab.io.vn/writing/proving-a-pipeline-is-replayable/https://tgblab.io.vn/writing/proving-a-pipeline-is-replayable/Replayability becomes credible when controlled failures produce independently observed evidence that named invariants still hold.Wed, 12 Aug 2026 08:18:00 GMTData systemsSoftware designbaokdl2226@gmail.comValidation Is Not a Publish Protocolhttps://tgblab.io.vn/writing/validation-is-not-a-publish-protocol/https://tgblab.io.vn/writing/validation-is-not-a-publish-protocol/Passing checks matters only when the evidence, candidate and reader-visible commit belong to the same revision.Wed, 12 Aug 2026 07:25:00 GMTData systemsSoftware designbaokdl2226@gmail.comLate Data Without Endless Reprocessinghttps://tgblab.io.vn/writing/late-data-without-endless-reprocessing/https://tgblab.io.vn/writing/late-data-without-endless-reprocessing/A correction window should bound automatic work without turning older truth into dropped data or a full-history rebuild.Wed, 12 Aug 2026 06:45:00 GMTData systemsSoftware designbaokdl2226@gmail.comIdempotent Writes Are Not Replay Safetyhttps://tgblab.io.vn/writing/idempotent-writes-are-not-replay-safety/https://tgblab.io.vn/writing/idempotent-writes-are-not-replay-safety/A table can converge perfectly while the workflow around it still repeats the wrong effect.Wed, 12 Aug 2026 05:30:00 GMTData systemsSoftware designbaokdl2226@gmail.comBackfills Are a Data Model Problem, Not an Airflow Featurehttps://tgblab.io.vn/writing/backfills-are-a-data-model-problem/https://tgblab.io.vn/writing/backfills-are-a-data-model-problem/A scheduler can recreate yesterday's run. Only the data model can make that run safe to publish again.Wed, 12 Aug 2026 00:00:00 GMTData systemsSoftware designbaokdl2226@gmail.com