24 March 2019
Profiler in Mongo
Profiler can be used to setup logging on the mongo instance. Events captured by the profiler: CRUD Administrative Operations Configuration Operations
Profiler can be used to setup logging on the mongo instance.
Events captured by the profiler:
*CRUD
*Administrative Operations
*Configuration Operations
If we check the profiling level method definition we can see how the profiling method is to be used.
> db.setProfilingLevel
function (level, options) {
if (level < 0 || level > 2) {
var errorText = “input level “ + level + “ is out of range [0..2]”;
var errorObject = new Error(errorText);
errorObject[‘dbSetProfilingException’] = errorText;
throw errorObject;
}
var cmd = {profile: level};
if (isNumber(options)) {
cmd.slowms = options;
} else {
cmd = Object.extend(cmd, options);
}
return assert.commandWorked(this._dbCommand(cmd));
}
db.commandHelp method can be used to know details about different methods used in mongo.For profile setup the details about the command can be found as below:
> db.commandHelp(“profile”)
help for: profile controls the behaviour of the performance profiler, the fraction of eligible operations which are sampled for logging/profiling, and the threshold duration at which ops become eligible. See http://docs.mongodb.org/manual/reference/command/profile
Setting up profiling on the databases as level 2 can be done as below:
> db.setProfilingLevel(2)
{ “was” : 0, “slowms” : 100, “sampleRate” : 1, “ok” : 1 }
or
> db.setProfilingLevel(1,3)
{ “was” : 2, “slowms” : 100, “sampleRate” : 1, “ok” : 1 }
The details about the profiling set can be seen by queryinig the system collection
> db.system.profile.find().sort({$natural:-1}).limit(1).pretty()
{
“op” : “query”,
“ns” : “config.system.profile”,
“command” : {
“find” : “system.profile”,
“filter” : {
},
“lsid” : {
“id” : UUID(“f563a6c1-6877-447b-93c5-59ff4bf75308”)
},
“$db” : “config”
},
“keysExamined” : 0,
“docsExamined” : 1,
“cursorExhausted” : true,
“numYield” : 0,
“nreturned” : 1,
“locks” : {
“Global” : {
“acquireCount” : {
“r” : NumberLong(1)
}
},
“Database” : {
“acquireCount” : {
“r” : NumberLong(1)
}
},
“Collection” : {
“acquireCount” : {
“r” : NumberLong(1)
}
}
},
“responseLength” : 1191,
“protocol” : “op_msg”,
“millis” : 0,
“planSummary” : “COLLSCAN”,
“execStats” : {
“stage” : “COLLSCAN”,
“nReturned” : 1,
“executionTimeMillisEstimate” : 0,
“works” : 3,
“advanced” : 1,
“needTime” : 1,
“needYield” : 0,
“saveState” : 0,
“restoreState” : 0,
“isEOF” : 1,
“invalidates” : 0,
“direction” : “forward”,
“docsExamined” : 1
},
“ts” : ISODate(“2019-03-23T15:37:27.842Z”),
“client” : “127.0.0.1”,
“appName” : “MongoDB Shell”,
“allUsers” : [ ],
“user” : “”
}
We can also see the last set of operations on the database by querying the system collection
> db.system.profile.find({},{op:1}).sort({$natural:-1}).limit(4).pretty()
{ “op” : “query” }
{ “op” : “update” }
{ “op” : “query” }
{ “op” : “query” }
We can also check details about the profiling using
>show profile
To check what level of profiling is set on the instance we can use
> db.getProfilingStatus()
{ “was” : 2, “slowms” : 100, “sampleRate” : 1 }
In Profiling data will be writing from beginning again , once it reaches end in profile collection.It has no indexes also, to speed up write to profile collection.