Minimal Always On Availability Group Design for Zero Data Loss in SQL Server
Learn the smallest viable Always On AG that guarantees zero data loss, its trust boundaries, key operational checks, and when to revisit the design.
Loading…
ReadMeFeed / Community knowledge
#always-on
Real questions. Useful conversations. Find the people who know your stack.
microsoftsqlserverChoosing Between Free Ephemeral and Paid Always-On Deployments on Replit
Decide whether to use Replit’s free ephemeral containers or paid Always‑On deployments based on budget, uptime SLA, latency tolerance, and storage needs, then validate the choice with a Flask health‑check and external monitoring.
replitAlways-On repl idle timeout behavior for idle TCP connections
Always‑On repls on Replit enter a sleep state after a period of no network activity, pausing CPU execution while preserving the filesystem. The platform does not expose the exact idle threshold, and its documentation does not clarify whether the timeout is triggered only when there is no data transfer on existing TCP connections or also when those connection
Always On – Lacking an Automatic Suspension Threshold
Always On – Lacking an Automatic Suspension Threshold Goal: Determine whether Replit can automatically suspend an "Always On" repl when traffic falls below a certain threshold to reduce costs for low‑traffic workloads. Constraint: Current documentation only allows manual enable/disable; no CPU or request‑based criteria for auto‑suspension. Uncertainty: Does
Choosing Between Glitch Pro Always‑on and External Ping for Keeping Containers Awake
Glitch Container Sleep and Timer Behavior The goal is to understand whether JavaScript timers (setInterval/setTimeout) continue to advance while a Glitch project is in its automatic sleep state after five minutes of inactivity. Constraints include the free tier’s 5‑minute sleep threshold, bandwidth limits for external ping services, and continuous RAM usage
glitch