I’m using overpass turbo, and trying to execute that query
[out:json];
{{geocodeArea:Germany}}->.searchArea;
way"power"=“line”;
out geom;
to get the power transmission lines in the Germany area, but it takes around 30 seconds to get executed, which is not suitable, so how to optimize that query just to return only the coordinates of each transmission line not meta data, and not tag names or details just coordinates, and make it faster to fetch the data.
That execution time seems appropriate, given that you are requesting a lot of data 
In JOSM:
[out:json][timeout:120];
{{geocodeArea:Germany}}->.searchArea;
way(area.searchArea)["power"="line"];
(._;>;); out meta;
Returns:
Even when saved as a .json with all tags removed, that is 52.6 MB!
because I’m running a real-time visualization engine that takes this data returned from the API, then runs some preprocessing on it and embeds it as KML, all of which takes less time than data fetching. try thinking of a country like USA, it would take forever !!!
So, I was thinking if there’s a way to reorganize the query or rewrite it in way to make it more optimum, in another way I want just coordinates of the lines not names nor tags or description, so it may decrease the request time.
What kind of real time changes to the power network do you need to display? Could you use periodic extracts rather than querying the live database via overpass?
I’m fetching the data, then extracting the coordinates, then appending them into the LineString tag in KML and visualizing them on Google Earth. The whole idea is that the user has to choose a country to visualize its electricity infrastructure, so it has to be kinda of real-time because of UX. Do you mean that periodic extract is used to extract all the data and use it as a local database that gets upgraded periodically?
It sounds a bit like OpenInfraMap. The About page has some details about how it is done, maybe it would give you some hints.
you can fetch this data from time to time (daily? monthly?) and display that data, rather than hit Overpass server whenever anyone opens your site
in other words, cache it
yes, though you may also keep it as a file or in other simple format
If you’re looking for fast queries of OpenStreetMap data, check out GeoDesk. You can turn any OSM PBF file into a GOL (a geographic object library – a compact single-file database stored on your own system). There are SDKs for Python, Java and C++, as well as a command-line utility – all free & open-source. A query to retrieve every w[power=line] in Germany should execute in less than 100 milliseconds on a modern machine.
GeoDesk doesn’t support KML as a native output format, but there are options for emitting OSM data as WKT or GeoJSON. Using the toolkits, it should be fairly trivial to write a converter that turns the Java/C++/Python objects returned by a GeoDesk query into the XML format of KML.
