开发者

Import RRD data into Python Data Structures

开发者 https://www.devze.com 2023-02-07 02:33 出处:网络
Does anyone have a good method for importing rrd data into python? The only libraries I have found so far are just wrappers for the command line or provide importing data into rrd and graphing it.

Does anyone have a good method for importing rrd data into python? The only libraries I have found so far are just wrappers for the command line or provide importing data into rrd and graphing it.

I am aware of the export and dump options of rrd, but I am wondering if开发者_如何学C someone has already done the heavy lifting here.


Here's an excerpt from a script that I wrote to get cacti rrd data out. It's not likely to be exactly what you want, but it might give you a good start. The intent of my script is to turn Cacti into a data warehouse, so I tend to pull out a lot of averages, max, or min data. I also have some flags for tossing upper or lower range spikes, multiplying in case I want to turn "bytes/sec" into something more usable, like "mb/hour"...

If you want exact one-to-one data copy, you might need to tweak this a bit.

    value_dict = {}
    for file in files:
        if file[0] == '':
            continue
        file = file[0]
        value_dict[file] = {}
        starttime = 0
        endtime = 0
        cmd = '%s fetch %s %s -s %s -e %s 2>&1' % (options.rrdtool, file, options.cf, options.start, options.end)
        if options.verbose: print cmd
        output = os.popen(cmd).readlines()
        dsources = output[0].split()
        if dsources[0].startswith('ERROR'):
            if options.verbose:
                print output[0]
            continue
        if not options.source:
            source = 0
        else:
            try:
                source = dsources.index(options.source)
            except:
                print "Invalid data source, options are: %s" % (dsources)
                sys.exit(0)

        data = output[3:]
        for val in data:
            val = val.split()
            time = int(val[0][:-1])
            val = float(val[source+1])
            # make sure it's not invalid numerical data, and also an actual number
            ok = 1
            if options.lowerrange:
                if val < options.lowerrange: ok = 0
            if options.upperrange:
                if val > options.upperrange: ok = 0
            if ((options.toss and val != options.toss and val == val) or val == val) and ok:
                if starttime == 0:
                    # this should be accurate for up to six months in the past
                    if options.start < -87000:
                        starttime = time - 1800
                    else:
                        starttime = time - 300
                else:
                    starttime = endtime
                endtime = time
                filehash[file] = 1
                val = val * options.multiply 
                values.append(val)
                value_dict[file][time] = val
                seconds = seconds + (endtime - starttime)
0

精彩评论

暂无评论...
验证码 换一张
取 消