Database monitoring is US$0–142 a month; 54% of outages cost over US$100,000

01Four meters, one free, and a US$142 ceiling

The database-monitoring market prices four different ways, and the unit matters more than the number. SolarWinds’ pricing page puts its Database category, the card behind Database Performance Analyzer and SQL Sentry, at “from US$142 per database a month” on multi-year contracts billed annually; neither product carries a list price of its own.¹ Datadog Database Monitoring is US$70 per database host a month billed annually, or US$84 on demand, with 200 normalised queries allotted per host, and it sits on top of an Infrastructure plan rather than replacing one.² Redgate Monitor Standard is licensed per server per year in bands of 1–4, 5–9 and 10–19 servers, however many instances a server runs, and 20 or more goes to sales; its own price table loads by script, and reseller listings put one server at US$1,164 a year, about US$97 a month.^{3 4} Percona Monitoring and Management is free under the AGPLv3 for MySQL, PostgreSQL, MongoDB and Valkey, with no paid edition.⁵

New Relic does not charge per host at all: 100 GB of ingest a month and one full-platform user free, then US$0.40 per GB plus US$49 a month for a core user or US$349 for a full-platform Pro user on an annual commitment.⁶ A quiet estate with twelve engineers who want dashboards is cheap on Datadog and expensive on New Relic; the reverse estate flips it.

02What the outage costs, from two surveys that name their method

Uptime Institute’s Annual Outage Analysis 2025, released on 6 May 2025, found that 54 per cent of respondents said their most recent significant outage cost more than US$100,000, 20 per cent said more than US$1 million, and only 9 per cent of incidents reported in 2024 were serious or severe.⁷ ITIC’s 2024 Hourly Cost of Downtime Survey, over 1,000 firms polled from November 2023 to mid-March 2024, found an hour of downtime costs more than US$300,000 for over 90 per cent of mid-size and large enterprises, and US$1 million to over US$5 million for 41 per cent, excluding litigation and penalties.⁸ Both are self-reported bands, not audited losses, and neither is India-specific; we found no Indian survey of comparable size and will not invent one.

03The mechanism is the same at every price

Every tool in the chart watches the same three things. Wait statistics say what sessions are waiting on: locks, disk, CPU, a network round trip. Slow-query capture ranks the statements that burned the most time in the last hour and keeps their execution plans, so a plan that flipped after an index rebuild shows up as a before-and-after. Replication lag says how far a replica is behind the primary, the number that decides whether a failover loses orders. An alert fires on a threshold or a learned baseline; a DBA kills the blocking session, adds the index or pauses the batch job; the user never sees the checkout stall. What you pay for above zero is the console, the engine list, hosting of the data, and a support line.