Performance tuning expert on TTFB tuning of PHP, apache2 & nginx under Ubuntu/Debian
Budget: $2 – $999 USD
We search for a performance in depth professional, wo is able to tune a production environment regarding php execution.
We need a real pro and engineer, who can
1) measure the current situation in terms of performance values E2E and especially on the server side, inside php, database, IO
2) share the report with your conclusions from the data
3) tweak the current configuration based on measured data
4) remeasure
5) verify if change has required impact
6) redo previous if impact is not visible
Infrastructure is available to measure internationally via servers distributed worldwide,
also we can measure locally in the LAN.
The browser based waterflow diagram (developer tools) can be shared as screenshots without NDA.
BUT if you only have this level of qualification to measure, you will not be successful.
gtmetrix, google pagespeed and other tools are also not helping here!
They will only proof our situation, which is:
The problem
Irregularly we have very long TTFBs for the first http-request of a website. (sometimes only few 100ms sometimes multiple seconds up to in worst case 10s)
we use mostly CMSes like wordpress, joomla and so on
The very first request takes ages in our case (sometimes multiple seconds) to deliver to the client
We need the optimization of the very first http request/response of opening the website, which contains all HTML and all resources loaded later.
we do NOT need a optimization (for now) of the later requests!
if you require to review code. See wordpress on github
What we do NOT need
- a plugin installer / deinstaller person
- a cache plugin person
- a CDN installer
- a simple sys admin, who installs some components and tries to find the issue by trial and error installations.
You are the clearly wrong person if you think you can narrow down by simply
- check free storage, ram, cpu usage , iops
- restarting services
- wanting to change/switching hosting provider (we own servers & infrastructure)
- better compression of images
- enabling gzip/br2 compression
- installing a CDN
- checking the slow query log
- checking gtmetrix, pagespeed, lighthouse, webpagetest and tuning the issues there (see waterfall section above)
=> either we did the checks or it will only coneal the real problem (e.g. CDN or cache plugins)
Such basic check are already done and also our monitoring is checking many of them
=> we need the optimization of the very first http request of opening the website
We will ask you in depth questions about the measuring in the qualification. So if you are not aware how to analyze the current situation.
You need to be clearly aware to be able to track/measure executions inside PHP to narrow down the problems.
Since we already get green values from measuring RAM, CPU, DB, Network, IO
But first browser http-request (TTFB) takes ages and afterwards the packages flow like rockets.
Can we switch provider or hardware or upgrade hardware?
provider-, hardware- and upgrade changes all tried. Incl. new setup on completely different provider, different cloud and newer ubuntu. No change.
So a clear NO, if you cannot explain with clean measurements E2E incl. all relevant system, network and software components.
Environment:
- dedicated hardware
- also vmware esxi
- apache webserver
- fpm-php
- nginx rev-proxy
- php 7.4.x / about to switch to 8.0 but no big benefit at all in tests
- Linux/Ubuntu
as a basic expertise proof we want to see a report showing this:
Δt_totalPhpExectime == Δt_databaseAccessTime + Δt_diskAccessTime + Δt_wpHookAccessTimes + Δt_wpPluginExecTimes + Δt_others
since getting lots of queries about URL of the page, this is what you will see
https://www.dropbox.com/s/4k5iv3vymjwpdzb/1stRequestTakesAges.png?dl=0
I am not interested in any later request tuning, since irrelevant. My issue is the marked and first request
Communication:
Do not wait for our availability here. Just answer, just ask or just reply.
Budget?
Not disclosed, place your best hourly rate to get considered
We need a real pro and engineer, who can
1) measure the current situation in terms of performance values E2E and especially on the server side, inside php, database, IO
2) share the report with your conclusions from the data
3) tweak the current configuration based on measured data
4) remeasure
5) verify if change has required impact
6) redo previous if impact is not visible
Infrastructure is available to measure internationally via servers distributed worldwide,
also we can measure locally in the LAN.
The browser based waterflow diagram (developer tools) can be shared as screenshots without NDA.
BUT if you only have this level of qualification to measure, you will not be successful.
gtmetrix, google pagespeed and other tools are also not helping here!
They will only proof our situation, which is:
The problem
Irregularly we have very long TTFBs for the first http-request of a website. (sometimes only few 100ms sometimes multiple seconds up to in worst case 10s)
we use mostly CMSes like wordpress, joomla and so on
The very first request takes ages in our case (sometimes multiple seconds) to deliver to the client
We need the optimization of the very first http request/response of opening the website, which contains all HTML and all resources loaded later.
we do NOT need a optimization (for now) of the later requests!
if you require to review code. See wordpress on github
What we do NOT need
- a plugin installer / deinstaller person
- a cache plugin person
- a CDN installer
- a simple sys admin, who installs some components and tries to find the issue by trial and error installations.
You are the clearly wrong person if you think you can narrow down by simply
- check free storage, ram, cpu usage , iops
- restarting services
- wanting to change/switching hosting provider (we own servers & infrastructure)
- better compression of images
- enabling gzip/br2 compression
- installing a CDN
- checking the slow query log
- checking gtmetrix, pagespeed, lighthouse, webpagetest and tuning the issues there (see waterfall section above)
=> either we did the checks or it will only coneal the real problem (e.g. CDN or cache plugins)
Such basic check are already done and also our monitoring is checking many of them
=> we need the optimization of the very first http request of opening the website
We will ask you in depth questions about the measuring in the qualification. So if you are not aware how to analyze the current situation.
You need to be clearly aware to be able to track/measure executions inside PHP to narrow down the problems.
Since we already get green values from measuring RAM, CPU, DB, Network, IO
But first browser http-request (TTFB) takes ages and afterwards the packages flow like rockets.
Can we switch provider or hardware or upgrade hardware?
provider-, hardware- and upgrade changes all tried. Incl. new setup on completely different provider, different cloud and newer ubuntu. No change.
So a clear NO, if you cannot explain with clean measurements E2E incl. all relevant system, network and software components.
Environment:
- dedicated hardware
- also vmware esxi
- apache webserver
- fpm-php
- nginx rev-proxy
- php 7.4.x / about to switch to 8.0 but no big benefit at all in tests
- Linux/Ubuntu
as a basic expertise proof we want to see a report showing this:
Δt_totalPhpExectime == Δt_databaseAccessTime + Δt_diskAccessTime + Δt_wpHookAccessTimes + Δt_wpPluginExecTimes + Δt_others
since getting lots of queries about URL of the page, this is what you will see
https://www.dropbox.com/s/4k5iv3vymjwpdzb/1stRequestTakesAges.png?dl=0
I am not interested in any later request tuning, since irrelevant. My issue is the marked and first request
Communication:
Do not wait for our availability here. Just answer, just ask or just reply.
Budget?
Not disclosed, place your best hourly rate to get considered