introduce "epoch" as a new base time reference, meaning timestamp 0.

[rrdtool.git] / doc / rrdfetch.pod

=head1 NAME

rrdfetch - Fetch data from an RRD.

=head1 SYNOPSIS

B<rrdtool> B<fetch> I<filename> I<CF>
S<[B<--resolution>|B<-r> I<resolution>]>
S<[B<--start>|B<-s> I<start>]>
S<[B<--end>|B<-e> I<end>]>
S<[B<--daemon> I<address>]>

=head1 DESCRIPTION

The B<fetch> function is normally used internally by the graph
function to get data from B<RRD>s. B<fetch> will analyze the B<RRD>
and try to retrieve the data in the resolution requested.
The data fetched is printed to stdout. I<*UNKNOWN*> data is often
represented by the string "NaN" depending on your OS's printf
function.

=over 8

=item I<filename>

the name of the B<RRD> you want to fetch the data from.

=item I<CF>

the consolidation function that is applied to the data you
want to fetch (AVERAGE,MIN,MAX,LAST)

=item B<--resolution>|B<-r> I<resolution> (default is the highest resolution)

the interval you want the values to have (seconds per
value). B<rrdfetch> will try to match your request, but it will return
data even if no absolute match is possible. B<NB.> See note below.

=item B<--start>|B<-s> I<start> (default end-1day)

start of the time series. A time in seconds since epoch (1970-01-01)
is required. Negative numbers are relative to the current time. By default,
one day worth of data will be fetched. See also AT-STYLE TIME SPECIFICATION
section for a detailed explanation on  ways to specify the start time.

=item B<--end>|B<-e> I<end> (default now)

the end of the time series in seconds since epoch. See also AT-STYLE
TIME SPECIFICATION section for a detailed explanation of how to
specify the end time.

=item B<--daemon> I<address>

Address of the L<rrdcached> daemon. If specified, a C<flush> command is sent
to the server before reading the RRD files. This allows B<rrdtool> to return
fresh data even if the daemon is configured to cache values for a long time.
For a list of accepted formats, see the B<-l> option in the L<rrdcached> manual.

 rrdtool fetch --daemon unix:/var/run/rrdcached.sock /var/lib/rrd/foo.rrd AVERAGE

Please note that due to thread-safety reasons, the time specified with B<-s>
and B<-e> cannot use the complex forms described in
L<"AT-STYLE TIME SPECIFICATION">. The only accepted arguments are "simple
integers". Positive values are interpreted as seconds since epoch, negative
values (and zero) are interpreted as relative to I<now>. So "1272535035" refers
to "09:57:15 (UCT), April 29th 2010" and "-3600" means "one hour ago".

=back

=head2 RESOLUTION INTERVAL

In order to get RRDtool to fetch anything other than the finest resolution RRA
B<both> the start and end time must be specified on boundaries that are
multiples of the desired resolution. Consider the following example:

 rrdtool create subdata.rrd -s 10 DS:ds0:GAUGE:300:0:U \
  RRA:AVERAGE:0.5:30:3600 \
  RRA:AVERAGE:0.5:90:1200 \
  RRA:AVERAGE:0.5:360:1200 \
  RRA:MAX:0.5:360:1200 \
  RRA:AVERAGE:0.5:8640:600 \
  RRA:MAX:0.5:8640:600

This RRD collects data every 10 seconds and stores its averages over 5
minutes, 15 minutes, 1 hour, and 1 day, as well as the maxima for 1 hour
and 1 day.

Consider now that you want to fetch the 15 minute average data for the
last hour.  You might try

 rrdtool fetch subdata.rrd AVERAGE -r 900 -s -1h

However, this will almost always result in a time series that is
B<NOT> in the 15 minute RRA. Therefore, the highest resolution RRA,
i.e. 5 minute averages, will be chosen which in this case is not
what you want.

Hence, make sure that

=over 3

=item 1.

both start and end time are a multiple of 900

=item 2.

both start and end time are within the desired RRA

=back

So, if time now is called "t", do

 end time == int(t/900)*900,
 start time == end time - 1hour,
 resolution == 900.

Using the bash shell, this could look be:

 TIME=$(date +%s)
 RRDRES=900
 rrdtool fetch subdata.rrd AVERAGE -r $RRDRES \
    -e $(($TIME/$RRDRES*$RRDRES)) -s e-1h

Or in Perl:

 perl -e '$ctime = time; $rrdres = 900; \
          system "rrdtool fetch subdata.rrd AVERAGE \
                  -r $rrdres -e @{[int($ctime/$rrdres)*$rrdres]} -s e-1h"'


=head2 AT-STYLE TIME SPECIFICATION

Apart from the traditional I<Seconds since epoch>, RRDtool does also
understand at-style time specification. The specification is called
"at-style" after the Unix command at(1) that has moderately complex
ways to specify time to run your job at a certain date and time. The
at-style specification consists of two parts: the B<TIME REFERENCE>
specification and the B<TIME OFFSET> specification.

=head2 TIME REFERENCE SPECIFICATION

The time reference specification is used, well, to establish a reference
moment in time (to which the time offset is then applied to). When present,
it should come first, when omitted, it defaults to B<now>. On its own part,
time reference consists of a I<time-of-day> reference (which should come
first, if present) and a I<day> reference.

The I<time-of-day> can be specified as B<HH:MM>, B<HH.MM>,
or just B<HH>. You can suffix it with B<am> or B<pm> or use
24-hours clock. Some special times of day are understood as well,
including B<midnight> (00:00), B<noon> (12:00) and British
B<teatime> (16:00).

The I<day> can be specified as I<month-name> I<day-of-the-month> and
optional a 2- or 4-digit I<year> number (e.g. March 8 1999). Alternatively,
you can use I<day-of-week-name> (e.g. Monday), or one of the words:
B<yesterday>, B<today>, B<tomorrow>. You can also specify the I<day> as a
full date in several numerical formats, including B<MM/DD/[YY]YY>,
B<DD.MM.[YY]YY>, or B<YYYYMMDD>.

I<NOTE1>: this is different from the original at(1) behavior, where a
single-number date is interpreted as MMDD[YY]YY.

I<NOTE2>: if you specify the I<day> in this way, the I<time-of-day> is
REQUIRED as well.

Finally, you can use the words B<now>, B<start>, B<end> or B<epoch> as your time
reference. B<Now> refers to the current moment (and is also the default
time reference). B<Start> (B<end>) can be used to specify a time
relative to the start (end) time for those tools that use these
categories (B<rrdfetch>, L<rrdgraph>) and B<epoch> indicates the 
*IX epoch (*IX timestamp 0 = 1970-01-01 00:00:00 UTC). B<epoch> is
useful to disambiguate between a timestamp value and some forms
of abbreviated date/time specifications, because it allows to use 
time offset specifications using units, eg. B<epoch>+19711205s unambiguously
denotes timestamp 19711205 and not 1971-12-05 00:00:00 UTC.

Month and day of the week names can be used in their naturally
abbreviated form (e.g., Dec for December, Sun for Sunday, etc.). The
words B<now>, B<start>, B<end> can be abbreviated as B<n>, B<s>, B<e>.

=head2 TIME OFFSET SPECIFICATION

The time offset specification is used to add/subtract certain time
intervals to/from the time reference moment. It consists of a I<sign>
(S<B<+> or B<->>) and an I<amount>. The following time units can be
used to specify the I<amount>: B<years>, B<months>, B<weeks>, B<days>,
B<hours>, B<minutes>, or B<seconds>. These units can be used in
singular or plural form, and abbreviated naturally or to a single
letter (e.g. +3days, -1wk, -3y). Several time units can be combined
(e.g., -5mon1w2d) or concatenated (e.g., -5h45min = -5h-45min =
-6h+15min = -7h+1h30m-15min, etc.)

I<NOTE3>: If you specify time offset in days, weeks, months, or years,
you will end with the time offset that may vary depending on your time
reference, because all those time units have no single well defined
time interval value (S<1 year> contains either 365 or 366 days, S<1 month>
is 28 to 31 days long, and even S<1 day> may be not equal to 24 hours
twice a year, when DST-related clock adjustments take place).
To cope with this, when you use days, weeks, months, or years
as your time offset units your time reference date is adjusted
accordingly without too much further effort to ensure anything
about it (in the hope that mktime(3) will take care of this later).
This may lead to some surprising (or even invalid!) results,
e.g. S<'May 31 -1month'> = S<'Apr 31'> (meaningless) = S<'May 1'>
(after mktime(3) normalization); in the EET timezone
'3:30am Mar 29 1999 -1 day' yields '3:30am Mar 28 1999' (Sunday)
which is an invalid time/date combination (because of 3am -> 4am DST
forward clock adjustment, see the below example).

In contrast, hours, minutes, and seconds are well defined time
intervals, and these are guaranteed to always produce time offsets
exactly as specified (e.g. for EET timezone, S<'8:00 Mar 27 1999 +2
days'> = S<'8:00 Mar 29 1999'>, but since there is 1-hour DST forward
clock adjustment that occurs around S<3:00 Mar 28 1999>, the actual
time interval between S<8:00 Mar 27 1999> and S<8:00 Mar 29 1999>
equals 47 hours; on the other hand, S<'8:00 Mar 27 1999 +48 hours'> =
S<'9:00 Mar 29 1999'>, as expected)

I<NOTE4>: The single-letter abbreviation for both B<months> and B<minutes>
is B<m>. To disambiguate them, the parser tries to read your S<mind :)>
by applying the following two heuristics:

=over 3

=item 1

If B<m> is used in context of (i.e. right after the) years,
months, weeks, or days it is assumed to mean B<months>, while
in the context of hours, minutes, and seconds it means minutes.
(e.g., in -1y6m or +3w1m B<m> is interpreted as B<months>, while in
-3h20m or +5s2m B<m> the parser decides for B<minutes>).

=item 2

Out of context (i.e. right after the B<+> or B<-> sign) the
meaning of B<m> is guessed from the number it directly follows.
Currently, if the number's absolute value is below 25 it is assumed
that B<m> means B<months>, otherwise it is treated as B<minutes>.
(e.g., -25m == -25 minutes, while +24m == +24 months)

=back

I<Final NOTES>: Time specification is case-insensitive.
Whitespace can be inserted freely or omitted altogether.
There are, however, cases when whitespace is required
(e.g., S<'midnight Thu'>). In this case you should either quote the
whole phrase to prevent it from being taken apart by your shell or use
'_' (underscore) or ',' (comma) which also count as whitespace
(e.g., midnight_Thu or midnight,Thu).


=head2 TIME SPECIFICATION EXAMPLES

I<Oct 12> -- October 12 this year

I<-1month> or I<-1m> -- current time of day, only a month before
(may yield surprises, see NOTE3 above).

I<noon yesterday -3hours> -- yesterday morning; can also be specified
as I<9am-1day>.

I<23:59 31.12.1999> -- 1 minute to the year 2000.

I<12/31/99 11:59pm> -- 1 minute to the year 2000 for imperialists.

I<12am 01/01/01> -- start of the new millennium

I<end-3weeks> or I<e-3w> -- 3 weeks before end time
(may be used as start time specification).

I<start+6hours> or I<s+6h> -- 6 hours after start time
(may be used as end time specification).

I<931225537> -- 18:45  July 5th, 1999
(yes, seconds since 1970 are valid as well).

I<19970703 12:45> -- 12:45  July 3th, 1997
(my favorite, and its even got an ISO number (8601)).

=head1 ENVIRONMENT VARIABLES

The following environment variables may be used to change the behavior of
C<rrdtoolE<nbsp>fetch>:

=over 4

=item B<RRDCACHED_ADDRESS>

If this environment variable is set it will have the same effect as specifying
the C<--daemon> option on the command line. If both are present, the command
line argument takes precedence.

=back

=head1 AUTHOR

Tobias Oetiker E<lt>tobi@oetiker.chE<gt>