Dynamic docker pool recycling
In one aspect, a system for dynamically recycling a pool of Docker containers is disclosed. The system includes a processor; a memory; and one or more modules stored in the memory and executable by a processor to perform operations. The operations include operations to manage for multiple accounts, monitored data received from agents installed at machines connected over a network, wherein the managed monitored data are associated with applications running over the machines for the multiple accounts; generate a model to estimates a Docker container pool capacity that can service the multiple accounts; receive performance alert action requests associated with the managed monitored data for at least a subset of the multiple accounts; based on the model, dynamically allocate a number of Docker containers in the pool that satisfies the estimated capacity; and update the model by performing simulations that uses additional data of Docker container usage.
1. A system for dynamically recycling a pool of Docker containers, the system including:
a processor;
a memory; and
one or more modules stored in the memory and executable by a processor to perform operations including:
manage for multiple accounts of a web service, monitored data received from agents installed at machines connected over a network, wherein the managed monitored data are associated with applications running over the machines for the multiple accounts;
receive performance alert action requests associated with the managed monitored data for at least a subset of the multiple accounts;
generate a model, the model configured to estimate a Docker container pool capacity that are available to service the multiple accounts, wherein the model includes a group of parameters associated with the Docker container recycling and the group of parameters includes stats on the alert action requests, wherein the stats on the alert action request include at least an average call volume and a spiking pattern for the alert action requests;
based on the model, dynamically allocate a number of Docker containers in the pool that satisfies the estimated capacity;
update the model by performing simulations that use additional data associated with Docker container usage; and
based on the updated model, dynamically recycle the Docker containers in the pool.
2. The system of claim 1 , wherein the group of parameters included in the modeled for Docker container recycling further includes Docker container launching time, Docker container killing time, or throughput.
3. The system of claim 1 , wherein the group of parameters included in the model for Docker container recycling further includes information on the recycling including whether the Docker containers are recycled immediately when being idle, recycled after certain delay, or recycled when needed.
4. The system of claim 1 , wherein the group of parameters included in the model for Docker container recycling is integrated with a scheduling algorithm.
5. The system of claim 1 , wherein the additional data includes historical data for Docker container usage obtained from representative accounts.
6. The system of claim 1 , wherein the one or more modules are executable to generate a pattern of alert action requests based on the determined stats.
7. The system of claim 6 , wherein the generated pattern of alert action requests is used to select the group of parameters for including in the model of Docker container recycling.
8. The system of claim 1 , wherein the one or more modules are executable to update the model including updating the stats on the alert action requests.
9. A method for dynamically recycling a pool of Docker containers, the method including:
managing for multiple accounts of a web service, monitored data received from agents installed at machines connected over a network, wherein the managed monitored data are associated with applications running over the machines for the multiple accounts;
receiving performance alert action requests associated with the managed monitored data for at least a subset of the multiple accounts;
generating a model, the model configured to estimate a Docker container pool capacity that are available to service the multiple accounts, wherein the model includes a group of parameters associated with the Docker container recycling and the group of parameters includes stats on the alert action requests, wherein the stats on the alert action request include at least an average call volume and a spiking pattern for the alert action requests;
based on the model, dynamically allocating a number of Docker containers in the pool that satisfies the estimated capacity;
updating the model by performing simulations that use additional data associated with Docker container usage; and
based on the updated model, dynamically recycling the Docker containers in the pool.
10. The method of claim 9 , the group of parameters included in the model for Docker container recycling is integrated with a scheduling algorithm.
11. The method of claim 9 , wherein the additional data includes historical data for Docker container usage obtained from representative accounts.
12. The method of claim 9 , wherein updating the model includes updating the stats on the alert action requests.
13. A non-transitory computer readable medium embodying instructions when executed by a processor to cause operations to be performed for dynamically recycling a pool of Docker containers, the operations including:
managing for multiple accounts of a web service, monitored data received from agents installed at machines connected over a network, wherein the managed monitored data are associated with applications running over the machines for the multiple accounts;
receiving performance alert action requests associated with the managed monitored data for at least a subset of the multiple accounts;
generating a model, the model configured to estimate a Docker container pool capacity that are available to service the multiple accounts, wherein the model includes a group of parameters associated with the Docker container recycling and the group of parameters includes stats on the alert action requests, wherein the stats on the alert action request include at least an average call volume and a spiking pattern for the alert action requests;
based on the model, dynamically allocating a number of Docker containers in the pool that satisfies the estimated capacity;
updating the model by performing simulations that use additional data associated with Docker container usage; and
based on the updated model, dynamically recycling the Docker containers in the pool.
14. The non-transitory computer readable medium of claim 13 , wherein the group of parameters modeled for Docker container recycling includes Docker container launching time, Docker container killing time, or throughput.
15. The non-transitory computer readable medium of claim 13 , wherein the group of parameters modeled for Docker container recycling further includes information on the recycling including whether the Docker containers are recycled immediately when being idle, recycled after certain delay, or recycled when needed.