Support for exact number of calls

[ldemailly]

Fixes #75
If you set RunnerOptions.Exactly to any number that exact number of
calls will be made, and without additional warmup - it will still try
to reach the target QPS but if the service is slower it will take the
time it will take

On the command line this mode is selected using “-n num”
(The replaces/extends the grpc ping count flag)

[codecov]

Codecov Report

Merging #82 into master will increase coverage by 0.8%.
The diff coverage is 91.3%.

@@           Coverage Diff            @@
##           master     #82     +/-   ##
========================================
+ Coverage    82.2%   82.9%   +0.8%     
========================================
  Files           6       6             
  Lines        1311    1313      +2     
========================================
+ Hits         1077    1089     +12     
+ Misses        164     155      -9     
+ Partials       70      69      -1

Impacted Files	Coverage Δ
periodic/periodic.go	97.4% <100%> (+0.3%)	⬆️
fgrpc/grpcrunner.go	75.8% <33.3%> (+5.6%)	⬆️
fhttp/httprunner.go	80.6% <66.7%> (+0.3%)	⬆️
fhttp/http.go	69.8% <0%> (+0.2%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 488e250...2c825f4. Read the comment docs.

[ldemailly]

@mandarjog this could help tracking down the rate limit and metrics issues:
if you ask it to do N requests it will do exactly N whatever the time it takes and without extra warmup calls

[mandarjog]

@ldemailly used it right away with -n 100, thanks.

[ldemailly]

@ldemailly used it right away with -n 100, thanks.

@mandarjog cool! I'm glad ! you can also use the ui if you run fortio server either inside or outside your cluster

And once istio/istio#2596 goes in, from test code, using

Exactly: 100, 

or o.Exactly = 100 in the (HTTP)RunnerOptions