Resource Usage / Background Writer

bgwriter_delay

• What it does: Sets the delay between rounds of activity for the background writer process, controlling how frequently it wakes up to write dirty buffers to disk.
• Why it matters: This parameter balances between timely writeback of dirty buffers and system resource usage. Shorter delays make the background writer more aggressive about writing dirty pages, which can reduce checkpoint overhead and smooth out I/O patterns. Longer delays reduce background writer activity but may lead to more intensive write activity during checkpoints.
• Ideal value & Best Practice: Default 200ms is reasonable for most workloads. For write-intensive systems, consider reducing to 100ms to make the background writer more aggressive. For read-heavy systems with few writes, increasing to 300-400ms can reduce unnecessary background writer activity.

bgwriter_flush_after

• What it does: Controls how frequently the background writer forces OS cache flushes by specifying the number of pages after which writes are flushed to persistent storage.
• Why it does: This parameter helps manage writeback behavior and durability guarantees. More frequent flushing ensures data is persisted to disk sooner but may impact performance. Less frequent flushing improves performance but increases the window for potential data loss. This works in conjunction with the OS and storage layer write caching.
• Ideal value & Best Practice: Default 512KB (64 pages of 8KB). For systems with battery-backed write cache or reliable storage, you can increase to 1-2MB for better performance. For systems without write protection, consider lower values like 256KB for better durability.

bgwriter_lru_maxpages

• What it does: Sets the maximum number of least-recently-used (LRU) buffers the background writer will write out per round.
• Why it matters: This parameter limits the amount of work the background writer does in each round, preventing it from consuming excessive I/O bandwidth and impacting foreground queries. It helps balance between keeping buffers clean and maintaining system responsiveness.
• Ideal value & Best Practice: Default 100 pages (800KB) per round. For systems with fast storage and high write throughput, increase to 200-400 pages. For systems with I/O constraints, decrease to 50-80 pages to limit background writer impact on foreground queries.

bgwriter_lru_multiplier

• What it does: Sets the multiplier used to calculate how many buffers the background writer should aim to keep ahead of current demand.
• Why it matters: This adaptive parameter helps the background writer anticipate future buffer needs based on recent usage patterns. Higher values make the background writer more aggressive about keeping free buffers available, which can improve performance by reducing the need for backend processes to perform their own writes. Lower values make the background writer more conservative.
• Ideal value & Best Practice: Default 2.0 is generally effective. Increase to 3.0-4.0 for systems with predictable, steady workloads. Decrease to 1.0-1.5 for highly variable workloads where aggressive background writing might be wasteful. Monitor buffer allocation statistics to tune this effectively.

---

Try pghealth Free Today 🚀

Start your journey toward a healthier PostgreSQL with pghealth.
You can explore all features immediately with a free trial — no installation required.

👉 Start Free Trial