A server can answer every ping while its database has been stopped for an hour. 'The server is up' and 'the work is happening' are two completely different statements - and only one of them matters to your users.
Basic monitoring stops at availability: is the machine reachable? Serious monitoring asks the next question - is the application on that machine actually doing its job? This page explains what database and application monitors in a tool like ManageEngine Applications Manager check beyond the ping, and why each check exists.
Three greens and one dead business. The ping was never going to tell you.
The first step beyond ping is the service check: is the database service (for example the SQL Server service) actually running, and do that automatically every few minutes. Services stop for boring reasons - a failed update, a crashed process, a maintenance restart somebody forgot to undo. A service monitor catches in minutes what users would otherwise report at 9 AM sharp, all at once.
A running service can still be useless - hung, deadlocked or starved. Application monitors go further and ask the database a small test query, timing the answer. That response time is the single most honest number in database monitoring: flat means healthy, creeping means trouble is loading, timeout means users are already suffering even though every infrastructure light is green.
None of these show up in a ping, a CPU graph or a service check. All three show up in a database monitor.
At the top layer, monitoring imitates a user: open the login page, run the key transaction, measure it end to end. Web and URL monitors do the simple version of this. For Java and .NET applications, the APM Insight agent (a separate add-on in Applications Manager) goes inside the code and shows which method or query consumed the time. Be honest about the ladder: each layer costs more setup than the one below. Most offices need layers 1–3 on their database and layer 4's simple version on their main application - and that already puts them ahead of 90% of setups.
Do not boil the ocean. Pick the single database your business cannot work without, add the service check, the test-query response time and the disk/log monitors, and set the alert discipline from the alerts page in this series. One deeply watched critical server protects you more than fifty servers watched at ping depth.
See the full live demo - adding servers, reading the dashboard and setting alerts, step by step.
▶️ Watch Full VideoBecause running and responding are different states. A deadlocked or disk-starved database service stays 'running' while answering nothing. Only an actual query proves the database can work.
The transaction log or database disk filling up. It grows quietly for weeks, then stops the database in one minute. A space trend alert turns that emergency into a scheduled cleanup.
No. Service checks, response-time queries, connection and space monitors are the agentless core. APM Insight adds code-level tracing inside Java and .NET applications - valuable for developers, optional for most admin setups.